China’s ninth annual World AI Conference (WAIC) was held from July 17–20 in Shanghai. Escalating points of tension in the U.S.-China AI competition, including tit-for-tat cybersecurity measures aimed at frontier models and U.S. labs’ reiterated emphasis of the risks posed by China’s AI ascendancy, meant this year’s gathering would be a crucial venue for Beijing to pitch an alternate model for global AI development. Top Chinese leader Xi Jinping’s decision to give the WAIC’s keynote address further emphasized the Chinese government’s focus on the issue. DigiChina invited a group of specialists to weigh in on the implications of Xi’s speech, a new initiative released by the Cyberspace Administration of China and the National Development and Reform Commission, the establishment of the Shanghai-based World AI Cooperation Organization, and the seemingly coordinated release of Moonshot AI’s Kimi K3 model.

REBECCA ARCESATI
Head of the Science, Technology and Innovation Program, Mercator Institute for China Studies

Xi chose his in-person debut at the World AI Conference (WAIC) to deliver a rare speech on AI, representing one of his most significant public interventions on AI to date. Previously, Xi’s richest speeches on AI had come from Politburo Study Sessions, and outsiders only got to see summaries. Here, he was addressing both the nation and the world.

What emerges starkly from his speech is the Chinese leadership’s vision, evolved and matured over a decade, of a world where humans find ways to collaborate with machine intelligence. “How to get along with thinking machines?” Xi asks shortly after beginning his speech. We find references to “integration,” “connectivity,” and “collaboration”—all familiar concepts in Chinese AI policy, reflecting the CCP’s high expectations for physical AI to boost growth, development, and societal wellbeing. Not surprisingly, agents and embodied AI were in the spotlight at this year’s WAIC.

A precondition for all this, in the CCP’s view, is that AI technology remains firmly under human control. “Controllability” (可控) has been a consistent leitmotif in Chinese AI policy, but what is meant by “control” can vary widely. (For example, it may refer to political control, cybersecurity, or the preference toward domestic technology.) What is new and more interesting is Xi’s explicit reference to “loss of control” (失控), probably best understood as the risk of AI systems running afoul. Xi had already sounded the alarm at a Politburo Study Session in January in a discussion of future technologies writ large. However, this is his first known mention of loss of control risks specific to AI—and it appeared in a public speech.

Still, it is not immediately clear from Xi’s WAIC speech, nor from other policy signals, that China’s leaders are concerned about alignment risks as understood by the international AI safety community. One cue that Xi might be concerned about safety flaws in model design lies in his distinction between “inherent and secondary risks” (内生和衍生风险).

Xi’s choice of words could indicate an intensified worry that AI systems might behave unpredictably. If we take an essay that appeared in the party’s top theory journal a day before Xi’s WAIC speech at face value, then the CCP would seem to be leaning toward a wider definition of human control that encompasses the preservation of human judgment, subjectivity, and power in the AI age. By contrast, current international (especially Western) AI safety debates emphasize existential risks from superintelligence.

Media reports suggest that some restrictions on access to Chinese open-source models could be on the horizon, yet it remains unclear whether this emerging discussion on AI export controls stems from genuine concern around alignment risks or reflects Beijing’s growing preoccupation with technology and know-how leakage. It will be key to watch whether top-level signals like Xi’s speech result in an acceleration of technical safety research—or a tighter regulatory grip on the industry.

SAMM SACKS
Senior Fellow, New America and the Johns Hopkins SAIS Institute for America, China, and the Future of Global Affairs

People from the mountains to the sea (人山人海) combined with 100°-plus temperatures in Shanghai humidity made for a sweat-drenched expo hall at this year’s World AI Conference last weekend. I was relieved to find some company booths handing out cold water bottles—as long as I scanned my WeChat QR code to follow them. The Moonshot AI booth looked relatively barebones compared to more gimmicky displays elsewhere. I never would have thought that this was the very same company whose announcement of the powerful Kimi K3 model coincided with the WAIC and sparked a firestorm among U.S. policymakers. What Moonshot had on offer was a chance to win free tokens—if you follow them on RedNote (Xiaohongshu). “This is the next DeepSeek moment!” a young woman staffing the booth told me excitedly.

Many of the thousands of visitors focused on the humanoid robots in the expo hall, but I spent some time at booths showcasing robot actuators. Actuators enable robots’ physical movement and weight-bearing, like joints in animals, and they come in all sizes. Some of the smallest are used for robotic fingers and hands, which are quite difficult to make well but go a long way to improve industrial robots’ ability to perform fine tasks. This is one part of how AI technologies might, as President Xi Jinping has charged, drive “prosperity.”

Leadership speeches took place in a separate building across the street from the public throngs in the expo hall. Here is where government officials and academics used careful language on stage, signalling broad ambitions and policy guidance that companies will interpret and may benefit from aligning with. The smaller sub-forums I attended delved deeper into the theme of how to govern AI agents—how, in other words, to assure the “security” Xi calls for alongside prosperity.

This year’s theme, agents, were touted as a major driver of productivity, but they also create immense challenges for security and safety. Should agents be given legal status, including liability? Should humanoid robots have ID cards from factory to end of life—similar to a national ID card for citizens but with more digits—so they can be monitored and assessed for safety and ethics in real time? Will agents from one country be able to interoperate with those from another, and what technical standards should govern those interactions? What should governments do about the fact that AI needs access to training data but agents strain existing data protection guardrails like consent, purpose limitation, and minimization?

I wandered to the mall in between these two massive conference halls seeking more air conditioning away from the throngs. In the basement food court, I encountered a light-hearted counterpoint to the international narrative of a U.S.-China battle for AI supremacy: An elderly woman smiled and put up her dukes, while a humanoid robot posed next to a boxing ring with gloves at the ready. It was a reminder that beneath the grand rhetoric of great power competition are real people trying to make their way in an unknown technology future.

KENDRA SCHAEFER
Director of Tech Policy Research, Trivium China

In his speech to the WAIC, Xi’s reiteration of China’s commitment to open-source AI was notable, particularly considering recentspeculation that Chinese companies won’t be able to keep open-sourcing their tools much longer, for two reasons. First, training frontier models is expensive. The bigger models get, the more revenue AI firms will need—making closed-source proprietary models ever more tempting as revenue generators for Chinese firms. Second, as open-source models grow more powerful, the risks of releasing them into the wild grow ever greater.

In his speech, Xi indicated that China isn’t ready to walk away from open source just yet, safety concerns notwithstanding. The global popularity of Chinese open-source tools presents a once-in-a-generation opportunity to cultivate global dependency on the Chinese stack, particularly in the absence of cheap, open U.S. alternatives.

However, he also made it very clear that attention to AI regulation, ethics, and governance has been elevated substantially within the Chinese government. As a result, we should expect a tightening of AI regulation in China. The speech calls for new AI laws and regulations, technical risk-monitoring systems, and emergency response mechanisms. In the immediate future, this will most likely result in ramped up pre-release capability testing and stricter rules on what Chinese AI providers can release.

Xi’s support for AI development, and open source in particular, goes hand in hand with his focus on stability. We can expect China to vigorously pursue AI competition, keeping both development and stability ahead of any near-term gains.

That competition has global valence. The founding of WAICO represents China’s bid to compete with the US on global AI governance (a deliberate counterweight to Western-led bodies like the Global Partnership on AI) and a pathway for pushing the Global South onto China’s AI stack. What’s interesting here is the bet Xi is making: that developing countries, priced out of frontier US models and wary of being rule-takers in institutions they didn’t build, will find a Shanghai-based alternative attractive. China’s efforts at establishing tech governance organizations haven’t historically been particularly successful, but this offer is more attractive than most. Still, I won’t be convinced the effort is gaining real traction until I see sign-on from a broader range of countries.

This Forum entry was adapted from Schaefer’s commentary originally published in the Trivium Tech Daily newsletter.

JOHANNA COSTIGAN
Editor-at-Large, DigiChina

The WAIC clarified that Beijing sees no need to be subtle in its attempt to recruit friends not from the west, but from the far larger rest. China’s National Development and Reform Commission’s AI Development Cooperation Action Plan includes a pointed line on “contributing global public scientific goods to the Global South.” The cyberspace administration’s new initiative on global governance of AI agents dedicates a section to “respecting diverse governance models.”

And in his keynote address, Xi Jinping said developing and deploying AI “should not erode or undermine the diversity of world civilizations or the cultural uniqueness of other nations.” Policymakers in Washington are primed to dismiss lines like this, but they are not the target audience.

China’s offers to protect developing countries from the threat and arguable reality of exclusion from the emerging AI-driven economy probably carry more weight to leaders (and citizens) of countries that are not home to frontier AI labs. At the risk of pointing out the obvious, that covers most of the world.

Still, China’s high-minded proposal of AI sovereignty and simultaneous equality will face severe obstacles in application. For one, Beijing’s stubborn claim that it is part of the developing world is increasingly strained as China’s geoeconomic power becomes harder to downplay. And nothing from these government documents is likely to convince anyone that China wouldn’t prioritize its political preferences over the promises of diversity and equality made ahead of the WAIC.

Take Taiwan’s approach to AI, for example. In an effort to reduce dependencies on Chinese training data and models, Taiwan is advancing a local LLM built on Taiwanese government data, Taiwanese ethnic minority languages, and traditional Chinese characters (the PRC uses a simplified script). It’s hard to imagine Beijing supporting that effort, no matter how many times it calls for respecting other cultures in AI development.

XIAOMENG LU
Director, Geo-Technology, Eurasia Group

One notable detail in Xi’s WAIC speech was his prominent placement of “inherent and secondary” (内生和衍生) AI risks—and the need for an early warning system. In the past, safety issues tended to stay at the bottom of leader commentaries, or they would receive a dutiful nod in a brief catch-all reference alongside other governance topics. A few days earlier, at a national science and technology conference, Xi also devoted more substantive attention to AI safety. Together, these signal a subtle but meaningful posture shift.

When Chinese models lagged, Chinese decision makers treated foundation model safety as a distant, and low priority concern. Sophisticated regulation of applications was deemed sufficient, and existential risk was not seen to be quite so imminent or material as some western researchers portray it to be. Now that Chinese models are approaching the frontier, Beijing’s rhetoric reflects a more forward-leaning posture on safety—a recognition that credible global leadership requires engaging with shared international concerns, in addition to showcasing capability.

Last year, over 200 prominent AI researchers and former heads of state signed the global Call for AI Red Lines, urging the UN to establish binding thresholds for AI development. As China’s models close the gap with their American competitors, its leaders have a greater incentive to demonstrate that they understand and share the concerns of international audiences such as those signatories—particularly in markets where adoption will have a meaningful impact on the global competitive landscape of AI.

Whether this shift is strategic or opportunistic, and whether words will translate into action, remains to be seen. For now, China emphasizes application safety and imposes modest requirements on foundation models. But this could represent a meaningful opening to revisit this issue in various forums, including the U.S.–China bilateral AI dialogue that is still taking shape. Comprehensive discussion on AI safety may be one of the few areas where American and Chinese interests begin to overlap.

Share

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu.

This DigiChina form was originally published at digichina.stanford.edu.

Share

Subscribe now

JUNMIN WANG
Professor of Sociology, University of Memphis

The Trump-Xi summit appeared unexciting and almost unfruitful on the core battlefield of science and technology (S&T) between the United States and China, with no major breakthroughs and only a few trade deals that were remotely relevant to S&T. The summit made incremental progress on the two governments’ mutual de-escalation efforts over the past six months, pulling further away from their dizzying, neck-to-neck confrontations in spring 2025.

With its continuously stringent control over high-end AI chips, advanced lithography equipment, and military technologies, the United States eased some export controls on AI chips (e.g., Nvidia’s H200) and deregulated trade in mid-range chips and equipment with China. For its part, China suspended export controls on rare earths and critical minerals until November 2026.

One of the summit’s profound yet less eye-catching achievements is the launch of institution-building efforts to address mutual vulnerabilities introduced by the U.S.–China trade and technology wars. The two countries agreed to establish the Board of Trade (贸易理事会) and the Board of Investment (投资理事会) to address future tensions and frictions, with the trade group reportedly to discuss tariff reductions on $30 billion of goods on each side. Institutionalizing negotiations in an effort to avoid “surprise attacks” (突然袭击), may not necessarily enhance trust or guarantee reliability. But it could be a good mechanism to prevent either side from arbitrarily destroying the current truce.

Emerging from the lowest point in the U.S.–China relationship in about half a century, the summit clearly focused less on immediate transactions and more on redefining the long-term relationship as both sides’ appetite for “decoupling” winds down. The two governments have agreed on a new concept for the U.S.–China relationship called “constructive strategic stability” (建设性战略稳定), another incremental update from the terminology of “strategic stability” that the U.S. government adopted in February 2026. Still, whether this “constructive” rhetoric can hold depends on whether the Trump administration maintains the pragmatism built over months leading up to the summit—and for how long.

LIZZI C. LEE
Fellow on Chinese Economy, Asia Society Policy Institute’s Center for China Analysis

The Trump-Xi summit offers a small but revealing glimpse into where Beijing’s tech policy may be headed. China will need to make progress on summit deliverables between now and Xi’s confirmed return visit to Washington in September. Two developments will be especially important.

First, will Beijing actually allow H200 deliveries, and if so, at what scale? The issue is not simply whether Washington approves the licenses. Previous openings around H200 access were effectively dead on arrival, because Beijing slow-walked purchases amid internal debate over whether such access would deepen dependence on U.S. technology. Has China’s domestic chip ecosystem developed enough since then that leaders feel confident H200 access will not hamper localization? Or has the compute shortage become so intense that Beijing is willing to tolerate more opening?

China’s top AI firms want the compute, especially for workloads where domestic capabilities still offer no real substitute. But China’s own security hawks remain suspicious and do not want U.S. chips to become anything more than a short-term bridge as China builds its own AI ecosystem. Actual shipments, volumes, and end-uses will show which bureaucratic camp is gaining ground in China: security hawks or development boosters.

Second, the AI safety protocol discussion is worth watching. Momentum in the days before the summit appears to have been at least partly driven by Anthropic’s limited release of Mythos, which pushed AI risk from abstract, feel-good governance language to a more muscular discussion about cyber vulnerabilities. China’s official response initially treated Mythos with skepticism, viewing its formidable capabilities as U.S. hype or a pretext for restricting Chinese access. But that skepticism has shifted toward concern that powerful agentic systems could create asymmetric risks for China itself. Beijing worries that U.S. firms could develop highly capable cyber technologies, restrict Chinese access to them, and still leave Chinese telecom networks, infrastructure, and logistics systems exposed. That gives China a real incentive to support an AI “best practices” protocol around misuse, cyberattack escalation, and crisis communication.

But Beijing also sees agentic AI as a leapfrog opportunity, so it will likely simultaneously pursue regulatory guardrails and accelerated innovation and deployment of agents. The tension between setting harder rules of the road and pushing faster innovation is now central to China’s AI policy.

AIDAN POWERS-RIGGS
Associate Fellow for China Analysis, iDeas Lab, Center for Strategic and International Studies

The Trump-Xi summit in Beijing underscored Taiwan’s paradoxical position in the AI era. President Xi remains fixated on dictating Taiwan’s political future, but the island’s growing centrality to global supply chains means much of the world is paying more attention to the material stakes of that future. President Trump, by contrast, has sent mixed signals on Taiwan, yet he cannot downgrade U.S. support for the island without imperiling the United States’ own AI ambitions.

TSMC is currently posting record revenues on the back of surging demand for AI chips. U.S. imports from Taiwan recently eclipsed those from China for the first time in decades. And despite geopolitical risks, the world’s leading technology firms continue to flock to the island, driving foreign direct investment to historic highs.

Taiwan’s economic success has thus become its best defense against political marginalization. Without buy-in from Taipei, any attempt to shape the island’s fate, whether from Beijing or Washington, would risk jeopardizing global economic stability.

The unresolved Nvidia H200 dispute shows how difficult it is for either side to escape Taiwan’s gravitational pull. The Trump administration has approved the chips’ sale to Chinese AI firms, but Beijing has so far held up the deal. Chinese leaders have prioritized technological self-reliance, but access to Nvidia’s AI chips would help ease compute constraints facing Chinese firms.

If the deal proceeds, it could accelerate parts of China’s AI ecosystem—but also further embed it within supply chains running through Taiwan. That may make the deal even more difficult for Beijing to accept; reliance on Nvidia can be written off as a temporary necessity, but depending on Taiwan is a reminder of the island’s rising status as an irreplaceable economic linchpin in the AI era.

CORY COMBS
Head of Supply Chain Research & Associate Director of Climate, Energy, Industrial Policy, Trivium China

From a summit stronger on vibes than solutions emerged one critical, concrete opportunity for deeper engagement: movement on the Board of Investment (BOI). For my money, it stands to be a defining mechanism for China-to-U.S. tech transfer.

More critically, the BOI would give the Department of Treasury more dominant influence in negotiations of where China can invest in the United States. It would shape what tech counts as “sensitive” in practice. To Treasury, national security is paramount. But it does not dismiss the potential economic benefits of Chinese investment and tech transfer quite as easily as other key agencies.

Treasury Secretary Bessent, who will chair the BOI, has explicitly said it would “make sure that these [non-sensitive] investments don’t get referred to CFIUS.” That’s significant assurance, given that uncertainty around whether or when the interagency Committee on Foreign Investment in the United States could unwind a deal remains a key holdup for massive amounts of potential U.S.-bound investment.

Plenty of investment focuses are obviously sensitive (biotech, AI) or nonsensitive (most traditional manufacturing, commercial properties not right next to military bases). But what matters is how determinations are made at the boundaries, where different interests argue in different directions. This is where it matters who is running the show.

Case in point: Whither batteries? For use in EVs made by American automakers? What about for industrial power supplies? And...AI data centers? Perhaps Chinese firms can license materials manufacturing technology and even set up factories, but will not get access to the electronics, software, or telemetry?

Private Chinese firms dominate the best lithium-ion technologies, the best emerging sodium-ion batteries, the best emerging hybrid batteries, and the United States is simply not positioned to catch up in the foreseeable future. The BOT is not a licensing mechanism, nor is it likely to clear Chinese frontier battery tech; the BOI is, and it could enable battery tech entry in ways made durable by genuinely addressing national security concerns.

In the meantime, the United States is quickly heading toward being an ICE (internal combustion engine) island—a nation of automakers uncompetitive in EVs and therefore uncompetitive abroad. This is recognized by even the most quintessentially American car giant, Ford, which is actively pursuing a partnership with Chinese battery leader CATL despite pockets of opposition from China hawks.

The United States has many such choices at hand, boiling down to: Where will it and where won’t it invite Chinese tech? How Treasury draws the lines of sensitivity via BOI on batteries and elsewhere will shape key areas of U.S. industrial competitiveness.

PAUL TRIOLO
Honorary Senior Fellow for Technology, Asia Society Policy Institute’s Center for China Analysis (CCA)

The central purpose of the summit was to reaffirm the October 2025 Busan pledge that neither side would take steps the other viewed as major escalations—especially U.S. export controls on semiconductors and equipment and Chinese controls on rare earths, magnets, and critical minerals. That did not appear to happen.

The lack of resolution is itself significant. Asked about formally extending the one-year trade truce reached last November, USTR Jamieson Greer said only, “We’ll see about that…There’s a willingness on both sides.” Meanwhile, U.S. companies already caught between the two countries’ export control regimes will continue to worry about long-term access to essential inputs. If agreement on this core issue slips to September, there are likely major unresolved disagreements inside both Washington and Beijing over the terms of any extension.

The larger question is whether both governments have absorbed the lesson that supply chain weaponization has been overdone. “Escalate to de-escalate” has become a dead end, generating collateral damage without creating stable leverage. Yet some in Washington still argue for more weaponization, assuming rare earth dependence can be solved quickly and painlessly. No and no. That mindset misunderstands the actual interdependence across the Pacific and risks pushing both countries into a cycle in which every control invites retaliation, and every retaliation expands the blast radius for private firms, allies, and global supply chains.

The summit also underscored the need for a real U.S.–China dialogue on frontier AI governance. Anthropic’s Mythos release has sharpened concerns in both capitals about models that could amplify cyber, biosecurity, and loss-of-control risks. Rather than pointing only toward more export controls, the Mythos moment inspires guardrails, crisis channels, and defensive cooperation. Washington should at least consider whether China can be brought under a protective umbrella for defensive uses, rather than leaving a major global actor outside shared mechanisms for hardening critical infrastructure.

In the end, a two-day summit was never going to resolve the structural problems in the relationship. It did not meaningfully strengthen the fragile détente—a goal that will remain elusive unless both presidents manage to restrain hardliners at home.

FEIFEI HUNG
Affiliated Researcher at the Asia Society Policy Institute’s Center for China Analysis (CCA)

As the Trump administration has approached Taiwan transactionally, evident in its use of arms sales as leverage and security commitments as bargaining chips with China, it is leaving Taiwan’s fate increasingly subject to the outcomes of bilateral U.S.–China negotiations it has no seat in.

The problem is compounding. If China is reducing its semiconductor dependency while the United States treats Taiwan’s security as negotiable, Taiwan’s room to maneuver is narrowing from both directions. The more pressing question for Taipei is not what was decided in Beijing, but what unique leverage it holds and how to deploy it.

Taiwan is not entirely passive. The government’s N-2 rule, which restricts TSMC from exporting its most advanced process nodes overseas, reflects an awareness that its technological edge requires active protection. But controlling the technology and having a say in how it is used geopolitically are different problems. Closing that gap carries its own risks.

Any move by Taipei that reads as hedging against Washington, particularly under a transactional Trump administration, could itself become a source of friction. Yet dependence on security guarantees from a partner that has shown willingness to negotiate over Taiwan’s core interests is not a stable posture either.

In light of Beijing’s increasingly aggressive stance on the Taiwan issue and Washington’s apparent ambivalence toward Taiwan’s security, it’s time for Taipei to enact more policies that defend its interests, even if such maneuvers anger one of both superpowers. If Taipei fails to act, it risks leaving itself vulnerable ahead of the upcoming season of U.S.–China summitry, including Xi’s planned autumn trip to Washington. It won’t have a seat at the stateside table either.

Share

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu.

This DigiChina Forum was edited by Johanna Costigan and Graham Webster. Check the web version of this page for further updates.

Note: Additional responses have been added since this Forum went live. Check the bottom for the latest!

Subscribe now

Share

KENDRA SCHAEFER
Partner, Trivium China

The 15th Five-Year Plan indicates that U.S. export controls on advanced semiconductors are pushing China to craft policies aimed at compute utilization efficiency as the country attempts to squeeze more output from the hardware it already has. The plan pursues this goal on three fronts: national infrastructure planning, local government oversight, and enterprise cloud adoption.

At the national level, the FYP calls to continue building the National Unified Computing Power Network and related state-backed scheduling platforms. The purpose of this network is to pool compute resources nationally and route workloads to underutilized data centers, ensuring that existing hardware is fully utilized.

At the local level, the plan moves to prevent local officials from wasting capital on ill-considered data center projects. In recent years, local governments raced to build data centers to hit regional digitalization targets and national KPIs, often without regard for local grid capacity, renewable energy availability, AI readiness, or actual demand. The FYP’s directive to “coordinate the layout and orderly construction of computing infrastructure” reasserts central planning authority over where capacity gets built.

At the enterprise level, the FYP calls to “actively develop public cloud services,” signaling a deliberate effort to migrate organizations onto hyperscaler platforms such as Alibaba Cloud and Huawei Cloud, another efficiency measure. State-backed think tank CAICT data places China’s cloud utilization rate at approximately 28%, against roughly 60% in the United States, a gap that reflects years of state-owned enterprises and government agencies defaulting to on-premises deployments. The result is scattered, fragmented data centers all competing for the same limited pool of AI-optimized semiconductors. These smaller installations cannot match the hyperscalers’ ability to efficiently allocate resources through greater expertise and advanced tooling.

Why the focus on wringing out waste? At this point, the strength of China’s high-end semiconductor manufacturing capabilities is an engineering problem. There is little more the state can do to accelerate R&D timelines. Infrastructure optimization, by contrast, is precisely where direct state intervention can move the needle. And regardless of how the chip export controls ultimately play out, a leaner, better-coordinated compute infrastructure will serve China well over the long term.

JEREMY WALLACE
A. Doak Barnett Professor of China Studies, Johns Hopkins School of Advanced International Studies

From a climate-oriented perspective, the 15th Five Year Plan is a disappointing, conservative document. To be sure, Chinese plans on clean energy have tended to be “under promise, over deliver” affairs. And there is the new Ecological and Environmental Code, which some see as laying the groundwork for a future Climate Law. Still, this document refuses to plan for the transition. Instead, it extends the current situation—where clean electricity covers all new demand but little or nothing is done to displace coal—for another five years.

It’s unclear to me why this conservatism is present. Of course, electricity grids are famously focused on reliability. Blackouts are serious political problems to be avoided, and memories of inadequate generation leading to shortages in 2021 are only five years old. Yet China has added a once-unfathomable amount of solar generating capacity over those five years. An additional 55 GW came online in 2021, then a conspicuously lucky 88 GW in 2022, followed by 216 GW, 278 GW, and 315 GW in the past three years, respectively.

That’s 950 GW in just solar. The country’s numerous solar manufacturers can now pump out 1,000 GW of panels every year. China could race to become the first electrostate, but instead, if the Five-Year Plan is any indication, it remains frozen.Even absent democratic decision making, it’s politically hard to lay off miners and cut off the tax-paying companies of the coal sector. It’s even harder when the mood of the country, especially its urban middle and upper classes, is demoralized by the policy-induced bursting of the real estate bubble. That being said, solar has come so far and so fast in the past five years that I dare to dream about what could be, despite a disappointing five-year plan. Perhaps the 15^th FYP will be eclipsed as dramatically as the 14^th FYP was in this space. After all, the green tech revolution is a huge mess that just might save the world.

ANGELA HUYUE ZHANG
Professor of Law, Gould School of Law, University of Southern California

China’s new Five-Year Plan offers a revealing answer to a question that observers often get wrong: What role does law play in Chinese technological development?

The conventional view is that China advances innovation through industrial policy and state subsidies, while law is largely irrelevant. But the 15th Five-Year Plan makes clear that this view is misleading. In fact, regulation features prominently in Chapter 5, which addresses emerging and future industries, including electric vehicles, advanced manufacturing, biopharmaceuticals, embodied AI, quantum technology, and brain-computer interfaces.

More specifically, the Plan explicitly calls for regulatory approaches including “sandbox regulation” and “trigger-based regulation,” as well as for legislation in emerging sectors such as biopharmaceuticals and “intelligent driving.” This points to a distinctive Chinese model of legal governance, one I have explored in a recent paper: Law should not be seen merely as a brake on technological change, but rather as critical infrastructure for innovation.

Consider first “trigger-based regulation.” The idea is straightforward. Instead of attempting to design comprehensive rules for every technology in advance, regulators allow innovation to move forward and intervene quickly when concrete risks emerge. In effect, the state tolerates uncertainty at the front end, often delaying sweeping legislative or regulatory intervention until problems become highly visible, systematic, and politically salient. There are certainly downsides to such an approach, as sharp regulatory swings can also spook investors and entrepreneurs, as seen in the previous tech crackdown. Nonetheless, permissive and tolerant regulation was central to the early growth of China’s internet economy, and a similar stance has now been applied to generative AI.

Now consider the other concept detailed in the Plan: “sandbox regulation.” Here the state does the opposite. In some sectors, innovation cannot proceed unless government acts first. Autonomous driving is the clearest example. Existing legal frameworks were not designed for fully driverless vehicles on public roads. If the law is not updated quickly, the technology remains trapped in the lab. So regulators have created bounded zones for experimentation, authorized pilot programs, and gradually constructed rules for liability, insurance, and technical standards. In those settings, law is not merely reacting to innovation; it is clearing a path for it.

This helps explain why China has moved so quickly in autonomous driving. Robotaxis, for instance, have now expanded to more than 20 Chinese cities, while commercial deployment in the United States has so far been permitted in only six, even though companies such as Waymo began experimenting years earlier.

The point is not that China is simply “better” at regulating technology. It is that China’s hierarchical political system can align legislation, administrative approval, and local experimentation more quickly when it decides that a sector is strategic. Democratic systems, by design, are more decentralized and often move more slowly. That can be a virtue, offering stronger protections for individual liberty and public safety. But it can also impose real coordination costs in fast-moving sectors.

Law, in other words, is not peripheral to China’s innovation model. It is central to it. The 15th Five-Year Plan makes that plain.

PAUL TRIOLO
Honorary Senior Fellow, Center for China Analysis, Asia Society Policy Institute

China’s 15th Five-Year Plan (2026–2030) marks a sharper, more security-driven technology agenda than its predecessor. The 14th Five-Year-Plan was drafted before Washington’s sweeping 2022–2025 semiconductor-related export controls and investment restrictions—and before generative AI became the defining field of technological competition.

As a result, the 14th FYP treated semiconductors, AI, quantum, and digital infrastructure as major development priorities, but kept them within a familiar framework of modernization and industrial upgrading. The new Plan outline is different in tone and design: it is much more explicit about breaking bottlenecks, securing “strategic initiative,” and reorganizing state support around resilient control of critical technologies and rapid deployment at scale.

What stands out most is that AI is no longer just one priority sector among many; it is the organizing logic for a broad industrial transformation. The new outline calls for progress on the comprehensive “AI+” action plan—building on implementation guidance released last summer. AI+ links AI to science, industry, culture, public services, and governance, and pairs AI development with a three-part buildout of computing power, algorithms, and data. It calls for national hub computing clusters, explores building ultra-large intelligent computing clusters, supports government purchasing of computing services and compute leasing, promotes coordinated siting of green electricity and compute, and aims to lower compute costs for small and medium-sized firms. It also pushes “model-chip-cloud-application” (模芯云用) coordination—while promoting multimodal models, agents, embodied intelligence, and even “explore” the development of artificial general intelligence. Just as important, it calls for national AI data corpora, sector-specific high-quality datasets, and a framework for the reasonable use of AI training data.

On semiconductors, the new plan is less about a single headline performance or capacity target than situating the chip ecosystem into a broader sovereign digital stack. The outline explicitly calls for upgrading high-end processors, optoelectronic devices, basic software, and industrial software, while highlighting the need for “decisive breakthroughs in key core technologies,” signaling unprecedented emphasis on self-reliance. In other words, Beijing is no longer thinking only in terms of catching up in one manufacturing node; it is trying to thicken the entire ecosystem around compute infrastructure, software, data, and application deployment under external pressure. This approach was also reflected in an important paper released just before the NPC by semiconductor industry leaders.

The plan is also more ambitious on robotics and future industries. It names robots, particularly humanoid versions, as among critical strategic industries—and elevates embodied intelligence to the same tier as quantum technology, brain-computer interfaces, 6G, and nuclear fusion. It also introduces structured mechanisms for the state to support future industries: large-scale application-demonstration programs, national AI application pilot and testing bases, concept-verification centers, and explicit risk-sharing mechanisms for investment.

Finally, the financing modes around government support for advanced technology is evolving. Major Chinese banks last week began ramping up lending to technology firms, with some targeting around 30% growth in loans to innovation-oriented companies; loans to small and medium tech firms had already risen to 3.63 trillion yuan by end-2025. That matters because the 15th Plan is not just prioritizing AI, semiconductors, and robotics rhetorically. It is building the financial plumbing—through banks, state planning, compute procurement, and industrial pilots—to fund them at scale in a far more hostile external technology environment.

Expect to see a lot more Chinese AI and advanced semiconductor and robotics firms tap capital markets in 2026, following on the heels of major IPOs in Shanghai and Hong Kong at the end of 2025 and early this year, including GPU players Biren, Moore Threads, Illuvatar, and Enflame, and AI model developers Zhipu, Minimax, and, soon, Moonshot. China’s five year plans have long since given up their planning aura in favor of high-level guidance, and the 15th iteration reflects both Beijing’s industrial priorities and its flexibility on how precisely they are achieved.

MARTIN CHORZEMPA
Dennis Weatherstone Senior Fellow, Peterson Institute for International Economics

China’s latest five-year plan reveals how central AI has become for China’s leadership. The plan mentions AI 52 times and includes a standalone AI+ action plan. By contrast, the last five-year plan from 2021 mentioned the term six times. AI has gone from a sideshow listed among other critical technologies to a key part of government ambitions for its economy and society.

China’s AI strategy can seem like the opposite of America’s: U.S. labs are framed as racing to artificial general intelligence (AGI) with closed source models, while Chinese are fast followers focused on diffusion of open models. The U.S. aims for limited regulation, while China organizes a massive industrial policy effort to develop indigenous capacity and reduce dependence on the United States for key semiconductors and equipment.

Yet, China’s approach has striking similarities to the U.S. AI Action Plan. Both frame AI as crucial to strategic competition. Both aim to accelerate AI development and diffusion, and both are positive on open-source AI despite safety risks.

The U.S. aims for “unquestioned and unchallenged global technological dominance.” China’s plan deploys military metaphors like “storm heavily fortified positions,” alongside a narrative of “win[ing] the strategic initiative” in “fierce international competition.”

Point after point in China’s plan calls to “accelerate” initiatives to develop and diffuse AI, reminiscent of the second Trump Administration’s support for AI “accelerationists” over more cautious voices accused of being “doomers.”

China’s government work report said that “China-made large AI models spearheaded the development of the global open-source AI ecosystem” and vowed to “support the development of open-source AI communities.” The U.S. plan more explicitly notes the “geostrategic value” of ensuring the United States has competitive open models to rival China’s. Notably, China’s plan does not preclude labs, such as Alibaba, going in a more closed direction.

Where the two countries differ the most is the role of the state. The U.S. plan is more focused on eliminating perceived barriers to AI development and adoption through deregulation. China’s more interventionist plan calls for the state to take on a more active role to foster “self-reliant, controllable” inputs for AI that do not depend on the United States. The plan thus includes efforts to make more advanced chips for training and inference, forming nationwide networks of computing power, and a national market for data—a longstanding ambition that is now more important as it could serve as a training data source for technical AI applications.

BEN MURPHY
Translation Manager, Center for Security and Emerging Technology (CSET), Georgetown University

One of the major problems the 15th Five-Year Plan is attempting to address is the innovation-application mismatch in China. Compared with competitors such as the United States, a much greater proportion of scientific research breakthroughs in China occur in state-run universities and research institutes rather than in corporate R&D labs. The disconnect between the academically oriented, state-funded researchers who make discoveries and the Chinese corporations and startups with the market knowledge and profit incentive to exploit innovations is a perpetual concern for policymakers.

Hence the 15th Five-Year Plan’s emphasis on the “mainstay status” (主体地位) of companies in technological innovation. The 14th Five-Year Plan (2020-2025) featured a lone call for “strengthening the status of enterprises as the principal entities of innovation,” and the fact that the 15th Five-Year Plan repeats similar language and elevates the phrase in importance (it is the subject of Chapter 10 of the Plan) is a tacit acknowledgement by China’s leaders that this problem has not been resolved and is in fact becoming worse.

And so—despite the irony that five-year plans are one of the most obvious holdovers from China’s Stalinist planned economy era—the 15th Five-Year Plan advocates putting more power and resources into the hands of Chinese companies. It calls for “increasing enterprises’ level of participation in national major technological innovation decision-making.” It encourages major corporate technology projects to be included, in an unspecified way, in state technology plans. It grants “greater autonomy” to “leading enterprises” in choosing which technical routes to pursue, which entities they collaborate with, and how they allocate expenditures when working on state-funded technology projects. A notable example of this policy in the 15th Five-Year Plan is the emphasis given to developing China’s commercial space sector, subject of a recent national plan of its own. Commercial spaceflight was mentioned only once in passing in the 14th Five-Year Plan, but in the 15th, it is one of 10 “new industries” to be developed, along with perennial priorities such as integrated circuits and large aircraft.

Of course, like many policies before it, the 15th Five-Year Plan is long on aspirations and short on specifics. Making more scientific data from state-funded projects available to companies—another proposal in the Plan—sounds smart, but who decides which companies get the data, and how much? If state-funded universities and labs are to permit “micro-, small-, and medium-sized enterprises” to use their innovations per the Plan, how will issues like licensing fees and IP ownership be handled? It remains to be seen whether and how policymakers will translate aspiration to reality.

MEI DANOWSKI
Co-founder and Principal, Natto Thoughts

China’s 15th Five-Year Plan (15th FYP) outlines a strategy based on integrating cybersecurity into every layer of digital development—from core infrastructure to international governance. This is not surprising, given the rapid development of cybersecurity in China.

Accelerating the construction of a “cyber superpower” (网络强国) is one of the five superpower-building areas highlighted in Part II of the 15th FYP. (The other four areas mentioned are: manufacturing superpower, quality superpower, aerospace superpower, and transportation superpower.) Furthermore, Part XIV discusses “Safeguarding National Economic Security.” The section titled “Enhancing Cybersecurity Capabilities” provides the main details of China’s strategy for cybersecurity and building a cyber superpower.

Particular emphasis is placed on protecting critical information infrastructure through cybersecurity reviews and cloud service assessments, as well as improving technical resilience by building disaster recovery and backup systems. There are also safeguards for industrial control systems and emerging technologies and applications. This is the first time that technical resilience—especially building disaster recovery and backup systems—has been included in the FYP. This likely reflects China’s recognition of the urgency amid growing technological competition and geopolitical tensions with Western nations.

At the same time, the 15th FYP promotes the growth of the cybersecurity industry as a pillar of economic and technological development. It calls for increased innovation, the production and adoption of secure and reliable information products and services, and the integration of cybersecurity into the modernization of manufacturing through intelligent transformation, digital transition, and network integration (智改数转网联). “Secure and reliable” products and services are aimed at reducing dependence on foreign products and services.

Beyond technical and industrial priorities, the 15th FYP underscores the importance of shaping a “healthy” cyberspace. Initiatives such as the “Qinglang” (Clean and Bright, 晴朗) campaigns, which first began in 2016, continue to curb online misinformation, cyber violence, and other content regulators disapprove of. Crackdowns on cybercrime, enhanced personal information protection, and tighter oversight of platform algorithms and data practices reinforce state control over the digital environment.

Overall, the cybersecurity strategy in the 15th FYP, compared to the 14th FYP, introduces more granular requirements, such as product standards for the development and use of “safe and reliable” information products and services, as well as technical resilience through disaster recovery and backup systems. There are also new security assessments specifically for cloud computing services. How these requirements are further specified and implemented may be worth continued attention over the next five years.

JOHANNA COSTIGAN
Independent Writer and Editor

Sticking to a timeline is especially difficult when you stand to gain from both meeting and missing your target. That’s the predicament the leaders of major AI labs in the U.S. have put themselves in when it comes to “achieving” artificial general intelligence (AGI). In February, OpenAI CEO Sam Altman described AGI as “a near-term thing.” In January, Anthropic’s Dario Amodei predicted an AGI equivalent (he dislikes the popular term) could come in the next 12 to 24 months.

In response to remarks like these, markets tilt and Washington stays gridlocked. Plenty of people disagree with such predictions, either rejecting the premise that AGI will ever appear, or debating when it will. Still, in America, the AGI conversation takes place largely in reaction to statements from the labs. In China, government intervention in AI discourse has resulted in more emphasis on the AIs of now relative to possible AGIs of the future. Officials have given companies the cues for what to focus on, packaged in policies and slogans such as the “AI+ Initiative” and the long-pursued “technological self-reliance.” The details of what to do with AI follow: upgrade existing industries, expand enterprise applications, innovate toward the efficient use of China’s limited compute, and adhere to national AI regulations.

The signaling on AGI has been tellingly sparse—yet, arguably equally tellingly, existent. Rather than completely ignore the term, Beijing has acknowledged it as a worthwhile pursuit for researchers without staking the future on AGI’s safe, functional, and assured emergence. This restrained approach was on display in both the 15th Five-Year Plan (FYP) and the 2026 Government Work Report.

The work report didn’t mention AGI. It focused on implementing China’s AI+ strategy and ensuring Chinese models lead the global open-source ecosystem, among other objectives. The FYP, an articulation of China’s big-picture policy strategy from now until 2030, featured AGI modestly, encouraging the “exploration of development pathways toward general-purpose AI.” That’s as close to a mention of AGI as the FYP gets. The term used, 通用人工智能, translates to “general-purpose” AI, applying the generality to the use case of the AI tool. That’s in contrast to the AGI definitions used in the U.S., which typically emphasize the human-like intelligence of the AI system. Including AGI in the FYP without pinning all hopes to it is prudent: rejecting the possibility of AGI is almost as extreme as insisting we’re on the brink of it.

Timelines are key to managing the AGI narrative tightrope, rendering Beijing’s time-locked policies useful. One thing is clear: Developing AI systems that are “general purpose” is of interest to policymakers. But in invoking “pathways,” they suggest this is a longer-term goal—one that will take a few years, during which the economic utility of other forms of AI should be thoroughly explored. It makes it to the five-year plan, but isn’t pressing enough to be included in the one-year plan.

The result is that China’s AI stakeholders, including the major model developers, investment community, and media ecosystem, know the boundaries. AGI can be discussed (DeepSeek CEO Liang Wenfeng, for example, has publicly stated his aspiration to create it). But government input clarifies that chasing AGI should not become China’s singular approach to AI development. That stands in contrast to the Silicon Valley-driven discourse, which frames the market and geopolitical competitions as fundamentally about AGI. China’s technocratic state is made of officials who are equal parts practical and radical: they forgo Silicon Valley utopia-speak, yet inject vast resources to science and technology “bets” that make every other government look cautious and passive. State planners are willing to note the potential of AGI without losing sight of the AI implementation and oversight that has to happen in the meantime.

CHARLES SUN
Yale Sinovation Fellow, Yale School of Management

Signals only matter if someone acts on them. What stands out about China’s approach to AI in 2026 is not the ambition of its agenda, but how quickly private firms have been moving to capture the economic advantages that agenda creates.

The plan calls for “government procurement of compute services” (政府购买算力服务), “cultivating a self-reliant and controllable software and hardware ecosystem” (自主可控的软硬件生态), and “greater government procurement of indigenously innovated products” (加大政府采购自主创新产品力度). Read together, these aspirations authorize a procurement and subsidy architecture that makes alignment with state objectives the cheapest market option. The Ministry of Finance has proposed a 20% discount for domestic products in government procurement evaluations. Energy subsidies have reportedly reduced power costs by up to half for some data centers deploying Chinese-made AI processors. Provincial and municipal governments have layered on additional incentives that consistently favor domestic hardware.

Firms are responding accordingly. Alibaba Cloud, a central player in Beijing’s push to migrate workloads onto domestic hyperscalers, has shipped hundreds of thousands of its in-house AI chips and built domestic-silicon clusters serving over 400 clients—because the subsidy structure makes its vertically integrated stack more cost-competitive. Zhipu AI optimized its GLM models for Huawei Ascend and Cambricon chips before the plan was even finalized, because compatibility with domestic hardware has become a de facto requirement for accessing hefty public-sector contracts. DeepSeek’s open-weight releases, most notably R1 under an MIT license, reflect a commercial bet that aligns with the plan’s support for open-source ecosystems: when compute is scarce, giving away the model layer and competing on efficiency is a rational strategy that happens to serve state goals.

The plan’s principle of “combining an effective market with a capable government” (有效市场和有为政府相结合) has appeared in five-year plans for years. In AI, it is becoming empirically visible in new ways. The state is not directing firms to build specific models or adopt particular chips. It is shaping the conditions under which commercial actors align with national priorities before they are formally required to do so. The incentives are local and tangible, and the cost of misreading direction is high. The more interesting question is not whether firms will align, but whether the AI bets embedded in China’s subsidy structure are technically sound, and what happens when they are not.

Subscribe now

Share

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu.

Beyond DeepSeek: China’s Diverse Open-Weight AI Ecosystem and Its Policy Implications

By Caroline Meinhardt, Sabina Nong, Graham Webster, Tatsunori Hashimoto, and Christopher D. Manning

After years of lagging behind, Chinese AI models – especially open-weight LLMs – seem to have caught up or even pulled ahead of their global counterparts. Models such as DeepSeek’s R1 or Alibaba’s Qwen family are now unavoidable in the global competitive landscape; they are widely adopted around the world and are performing at near state-of-the-art levels across major benchmarks.

In this brief, almost one year after DeepSeek’s attention-grabbing releases, we pause to give context to the technical and commercial realities of leading Chinese AI labs against a constantly evolving geopolitical backdrop. We profile and compare the capabilities and distinct features of four notable Chinese open-weight language model families, highlighting that China’s ecosystem of open-weight LLMs is driven by a wide range of actors who are prioritizing the development of computationally efficient models optimized for flexible downstream deployment. Finally, we highlight the policy implications of Chinese open-weight model diffusion, exploring how they may shape global technology access and reliance patterns, as well as their implications for AI governance, safety, and competition.

Addressing these developments will require policymakers to ground their policy actions in a more granular understanding of real-world deployments of open-weight AI models.

Read the brief:

Share

Subscribe now

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu.

Three months after the Chinese AI company DeepSeek shocked global markets with a highly capable reasoning model, another China-linked company made a splash with a capable agentic AI system. Did Manus, released in March 2025, portend Chinese leadership in AI systems that go beyond chatbots to take action on the user’s behalf? Victor Mustar, head of product at Hugging Face described Manus’ capabilities as “mind-blowing, redefining what’s possible.” A journalist’s comparison with ChatGPT DeepResearch found that Manus provided better results, despite speed and stability issues.

Manus had been released by a Singapore-based firm but developed by a startup in Wuhan with backing from the Chinese tech giant Tencent. It wasn’t China’s only foray into the emerging field. The same month, the Beijing-based firm Zhipu AI launched AutoGLM-Rumination, an open-source agentic system the company said achieved “state-of-the-art” scores on benchmarks such as AgentBench. (Zhipu also announced an “international alliance” for autonomous AI models, to include 10 countries associated with the Belt and Road Initiative and from ASEAN.) Earlier in January, Alibaba released the Qwen-Agent framework for building agentic systems with its Qwen models. ByteDance followed with its Coze Studio platform in July. Last month, Tencent open-sourced Youtu-Agent agentic framework, which was reportedly built atop a DeepSeek model.

With so much action this year in Chinese “agentic” AI efforts, it’s worth pausing to ask what Chinese developers mean when they talk about agentic AI. Moreover, what does the proliferation of such systems in China mean for AI safety and governance in the country?

Share

Lexicon: What we talk about when we talk about ‘agentic’

While “agentic AI” is a widely recognized phrase in English-language discussions of AI systems with the ability to perform complex actions, there is some inconsistency in how the technology is discussed in Chinese. The most common terms can be split into two groups:

• First, there are those that render “agent” or “agentic” using the word 代理 dàilǐ, which can mean “agent” in the sense that a dàilǐ can serve as a representative or proxy. It connotes action on someone’s behalf. Dàilǐ is also a verb meaning to act as an agent or representative of someone else. These forms include 代理型人工智能 dàilǐxíng réngōngzhìnéng or 代理式AI dàilǐshì AI (Agent-type AI) and AI代理 AI dàilǐ (AI agents).

• Second, there are those that emphasize the autonomy or independence of the AI system as an actor in the world. These include 自主代理 zìzhǔ dàilǐ (autonomous agent), 自主智能体 zìzhǔ zhìnéngtǐ (autonomous intelligent entity), and 智能体 zhìnéngtǐ(intelligent entity, or an intelligence).

Chinese government publications such as a new standard on AI agentsreleased on May 27 by the China Academy of Information and Communications Technology (CAICT) use the term 智能代理(intelligent agents). An earlier 2024 report on LLM benchmarking by the CAICT uses both 智能体 (intelligent entity) and 智能代理 (intelligent agent) and sometimes adds on a translation, clarifying how usage compares with English, e.g. “智能体(AGENT).” In academic journal articles, common terms include 大模型智能体 dàmóxíng zhìnéngtǐ (large model intelligent agents), 智能体 (intelligent entity),and 智能代理 (intelligent agent). Overall, the concurrent use of several different terms across Chinese-language discourse indicates that the discursive definition of an agentic AI system is still underway as the technology matures.

Share

Soft law and policy discourse

The regulatory context for these various systems is evolving. Some aspects of what might be considered agentic AI systems are already regulated in Chinese law. If they constitute recommendation algorithms, generative AI systems, or otherwise do things the government regulates regardless of the underlying technology, they are already regulated. When it comes to direct references to agentic systems, the action so far has come in the form of standards or specifications, sometimes considered “soft law.” Though such documents are generally nonbinding unless specified, when published by government think tanks, they can act as valuable guidance for businesses navigating an evolving regulatory space.

• On March 27, 2025, CAICT, in collaboration with more than 20 Chinese AI companies including Tencent, Alibaba, and Huawei, reportedly released a standard entitled “Technical and Application Requirements for Intelligent Agents in Software Engineering – Part 1: Development of Intelligent Agents” (《软件工程智能代理技术与应用要求 - 第1部分：智能代理的开发》). The standard reportedly focuses on capacity-building and application requirements for AI agent development, particularly regarding the technical and service capabilities of models. The release of this standard was heralded as a “new phase” in R&D and application of AI agents as part of a ramping up of AI agent commercialization in China and globally. The full text does not appear to be publicly available.

• The same week, the International Telecommunication Union Telecommunication Standardization Sector (ITU-T) approved an international recommendation on AI agents, reported by Chinese state media as spearheaded by CAICT. The ITU recommendation, titled “ITU-T F.748.46: Requirements and Evaluation Methods of Artificial Intelligence Agents Based on Large Scale Pre-Trained Model,” focuses on standards for evaluations that assess the capabilities of AI agents in tasks like perception, planning, execution, and collaborative interaction to enhance performance, both for single-agent and multi-agent systems.

• CAICT is also a center of China’s published policy thinking on agentic AI. In 2023, its annual “Blue Book Report on Large Model Governance” offered foundational guidance on assessing large-scale AI models and mentioned AI agents as part of a larger category of decision-making AI that should be regulated.

• In June 2024, CAICT published the Large Model Benchmarking System Research Report (2024), which discussed AI agents as part of a broader effort to introduce national benchmarking standards for the Chinese AI industry. In particular, the report raised the need for benchmarks to incorporate measures suited to assessing AI agents through the construction of evaluation environments. Building on this, the Fang Sheng (方升) benchmarking system introduced in the report placed a strong focus on “application-oriented testing” (大模型的应用测试, AOT) of large models to test challenges in their deployment in business settings. This includes considerations on how integration of large models with external tools to perform more complex tasks will require the use of AI agents.

• In December 2024, CAICT published the Blue Book on Artificial Intelligence Governance (2024), which, while not mentioning AI agents explicitly, does include broader reflections on the ethical implications of human-AI relationships, including who will be held responsible for decisions taken by AI systems and the impact emotional dependency on AI may have on human autonomy and relationships. The report also includes governance recommendations for full-lifecycle (全生命周期) oversight of AI systems across R&D, regulation, and third-party evaluation bodies, which can also apply to agentic AI systems.

• In 2025, the debate on agentic AI in China aligns with a broader focus shift towards AI competition and large-scale AI deployment across industries. This policy direction is perhaps best captured in Xi Jinping’s address on April 26, 2025, during the second Politburo study session on AI, which reaffirms China’s leadership aspirations in AI and emphasizes self-reliance and application-oriented development. However, it also highlights the role of the government in ensuring AI is “safe/secure, reliable, and controllable,” by “speed[ing] up the formulation and improvement of the relevant laws and regulations, policy system, application specifications, and ethical criteria, construct[ing] systems for technology monitoring, early warning for risks, and emergency response,” etc.

However one defines it, Chinese technological and policy developments in agentic or autonomous AI are advancing rapidly. Concrete regulation of this class of technologies remains embedded in preexisting legal and regulatory systems around data governance, cybersecurity, recommendation algorithms, and generative AI. The policy thinking embodied in CAICT’s standard and reports, meanwhile, points to the potential for governance to evolve with the technology’s application.

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu.

China’s top leader Xi Jinping last week presided over a Politburo study session on artificial intelligence. It was the second such AI-focused study session, the previous dating back to 2018, when Xi had just started his second five-year term and China’s government had recently published the New Generation AI Development plan, setting ambitious targets for AI development and governance as far out as 2030. Like in 2018, the most extensive public account of the session and Xi’s remarks comes in the form of a Xinhua report, which has been translated by Ben Murphy of the Center for Security and Emerging Technology (CSET) at Georgetown University.

Xi’s reported remarks echoed some of the party-state’s longstanding thinking on AI, while also reflecting today’s context: the Chinese startup DeepSeek’s impressive releases in December and January, more than two years of US and allied semiconductor controls designed to slow China’s advances, and intense economic pressures both at home and abroad. How should we understand the importance of this event and the messages Xi and the Politburo are sending? What does the carefully crafted language of the official media release suggest about their priorities and preoccupations?

DigiChina invited a group of specialists to share their insights on these questions. This Forum may be updated as additional contributions come in.

Note: Some contributors below adopt their own translations for quoted terms and passages, differing from CSET. I’ve attempted to flag substantive differences in footnotes.

–Graham Webster, Editor-in-Chief, DigiChina

Subscribe now

Share

KRISTY LOKE
Independent researcher

There are three key takeaways from China’s recent Politburo study session on “Strengthening AI Development and Governance”:

First, regarding China’s AI strategy, the study session confirms that Xi remains highly committed to the “new whole-of-nation” approach.^[1] The approach balances (a) strong organizational and resource support from the government with (b) the acknowledgement that it is the market, not the government, that can effectively deliver core science and technology (S&T) breakthroughs. This shows that DeepSeek’s success has not led to a strategic adjustment and that we can expect the government to play a pivotal but measured role in supporting AI development.

Second, regarding China’s core AI goals, the study session calls attention to AI adoption and applications. This is highly consistent with the Four Orientations (四个面向) concept introduced by Xi in 2020, which proposed that economic transformation and value creation should guide China’s S&T priorities. In other words, China views its competition with the United States as one that can be won via AI adoption, instead of a race toward the elusive artificial general intelligence (AGI). This emphasis on tech diffusion helps explain why, following the release of ChatGPT, the Chinese government chooses to focus on market and S&T reform, and improving the accessibility of compute and data nationwide, instead of placing key resources behind handpicked AGI champions.

Third, regarding AI governance, the study session sets the expectation that AI development should be “beneficial, safe, and fair”, with Xi reportedly emphasizing that “artificial intelligence can be a global public good that benefits mankind.” This reflects China’s growing commitment to supporting the Global South via tech capacity building and sharing governance-related know-how. It also echoes China’s long-term emphasis on applying a human-centric approach to AI development and the Politburo’s post-ChatGPT call on society to balance AI development with governance.

PAUL TRIOLO
Honorary Senior Fellow, Center for China Analysis, Asia Society Policy Institute

Politburo study sessions are important opportunities for experts to inform senior leaders about a particularly critical sector or topic, and for Chinese leaders, and in particular Party General Secretary Xi Jinping, to signal an industry on his views of the way forward on a particular issue, and also to signal the bureaucracy about his preferences.

At the 9th study session in 2018, in a much different world, including around AI and US–China relations, Xi called for the healthy development of AI. The topics of AI and semiconductors have featured in a number of the study sessions over the past decade, highlighting how close the senior leadership is to these issues.

The timing of this particular session is important, coming just a few months after the DeepSeek effect has swept China’s AI sector, with Xi and other Chinese leaders all paying close attention to the issue of AI development and increasingly concerned about US technology controls, domestic capabilities to overcome those controls, and the increasing sense that the US intends to race towards AGI with the goal of gaining a decisive strategic advantage over China and using it.

Xi’s call to “concentrate our efforts on mastering core technologies such as high-end microchips and foundational software and construct an independent/indigenous and controllable (自主可控)^[2] AI foundational software and hardware system with collaborative operations” is a direct response to what the Chinese leadership likely sees as an existential threat from falling behind on AI development, and the challenge posed by US policies such as the upcoming AI Diffusion Framework, which will attempt to isolate China from global AI development and give the US government a veto over where advanced AI capabilities are developed and deployed. For Xi and the Politburo, game on for AI supremacy.

HELEN TONER
Director of Strategy and Foundational Research Grants, Center for Security and Emerging Technology, Georgetown University

What I find most notable about this readout is how consistent it is with past CCP messaging on AI, despite the many changes in the underlying technology. While it does mention the "rapid progress" that has characterized the field of AI recently, there do not seem to be too many updates on how the Politburo is thinking about AI's challenges and opportunities. There is no mention of major developments in recent years (such as generative AI, large language models, reasoning models, agents, general-purpose AI), nor of potential future developments that are becoming common features of Silicon Valley discourse (such as AGI or superintelligence). The choice of expert speaker reinforces that these developments do not seem to be of special interest to the Politburo: As noted by Concordia AI, Prof. Zheng Nanning is an old-school machine learning researcher focused on computer vision and pattern recognition, not in the areas that are currently booming.

This is only one brief glimpse into the CCP's thinking on AI, and we know there has been a flurry of activity on creating standards and regulations for generative AI. Clearly they are not blind to recent developments. But it appears that at least to an external audience—and perhaps in their own thinking as well—they don’t want to center these newer issues, even while reiterating what a high priority AI is for the CCP's top leadership.

The other element that struck me was the prominent billing given to international engagement and Global South capacity-building, with the whole final paragraph dedicated to these issues. It's a safe bet that with the United States retreating from its international leadership role, this theme will continue to pervade China's AI policy.

JOHANNA COSTIGAN
Associate Director of Research, Paulson Institute

“We must use AI to lead a paradigm shift in scientific research and speed up S&T innovation and breakthroughs in every field,” Xi advised. He wants Chinese researchers to unleash the next “paradigm shift”—a term that has long been used to describe AI capability leaps among foreign AI researchers and appears to be gaining traction within the CCP.

For example, it was featured in multiple distinct contexts in a People’s Daily article published in mid-March. The first comports with the highly sci-fi way the AI industry talks about paradigm shifts, as redefining “the boundaries of human wisdom.” But a subsequent mention applies an AI "paradigm shift" to upgrading traditional industries—a top policy priority in itself that aligns with the readout’s emphasis on AI’s real-word deployment. A third reference shifts into the modalities of AI systems themselves.

Beyond industrial applications, which occupy a special place in policymakers’ minds, it appears paradigm-shift thinking has broad resonance in the post-DeepSeek era. Recently, for example, it was used to describe changes in academic research and the forthcoming transformation in administrative law from “fuzzy" judgments to “scientific and intelligent decision-making.”

Beijing’s conception of an AI paradigm shift, then, appears to encompass the version touted by figures like Dario Amodei and Sam Altman, who reference new techniques behind “reasoning” models. But China’s usage also emphasizes applications in scientific breakthroughs, harnessing AI for economic development, and creating relevant laws and regulations.

Contrary to assumptions that China won’t develop AGI because the CCP wouldn’t want or allow an uncontrollable superhuman intelligence, policymakers appear tacitly supportive, as long as whatever ends up being called AGI has economic utility and basic safety mechanisms. Rather than quashing AGI out of a fear of controllability, Beijing is more likely to both promote the A(G)I “paradigm shift” and seek out innovative ways to govern it.

SCOTT SINGER
Visiting Scholar, Carnegie Endowment for International Peace

Xi Jinping's recent Politburo study session on AI reinforces China's post-DeepSeek moment strategy: accelerating domestic AI adoption while pursuing technological self-sufficiency across the entire AI stack. The session underscores China's view of AI as a strategic technology and critical engine for economic growth amid domestic and international pressures.

What stands out against this backdrop of technological acceleration is Xi's especially strong statement on AI risks. His description of "unprecedented risks and challenges" marks his most direct acknowledgment of AI safety concerns to date. The speech outlines specific mitigation measures, calling for "technical monitoring, risk warning, and emergency response systems" alongside accelerated development of "relevant laws, regulations, policy systems, application specifications, and ethical guidelines." These concrete directives signal imminent regulatory action and suggest that while Chinese leadership views AI as an economic opportunity, it also recognizes the need to address substantial risks from powerful AI models. This apparent shift lends credibility to the China AI Safety and Development Association, indicating that China's self-proclaimed AI Safety Institute–equivalent is gaining meaningful influence.

Xi's adoption of the "global public good" framing—terminology directly borrowed from international statements like the 2024 Venice Statement at the International Dialogue on AI Safety—demonstrates China's engagement with global AI safety discourse. This language suggests that internationally developed concepts can be integrated into China's domestic context, potentially creating narrow but viable pathways for developing shared norms and standards in AI governance, even as technological competition intensifies. Collectively, these developments suggest that China's AI safety movement, though still a minority position in Chinese tech policy circles, is gaining meaningful momentum and influence domestically.

GABRIEL WAGNER
Affiliate, Concordia AI; Yenching Scholar

and JASON ZHOU
Senior Research Manager, Concordia AI

In his speech, Xi Jinping warns that while AI brings “unprecedented development opportunities,” it also poses “unprecedented risks and challenges.” This is the first time the Chinese leadership has described AI risks as “unprecedented,” suggesting elevated concern. Yet mentions of AI safety are not new. As early as the 2018 Politburo study session, leaders called for ensuring AI is “safe, reliable, and controllable”—a phrase Xi reiterated in 2025.^[1] The 2024 Third Plenum similarly emphasized the need to “institute oversight systems to ensure AI safety.”

What sets the 2025 session apart is both the language and the level of detail. The readout includes an entire paragraph focused on safety—unusual for such high-level documents, where safety is often addressed in a sentence or two. It makes a concrete call for “technology monitoring, early risk warning, and emergency response” systems. While this specific phrase is new, 2021 New Generation AI Ethics Guidelines (zh, en) and the February 2025 revision of the National Emergency Response Plan reference similar mechanisms.

Xi also urged faster development of “laws and regulations, policies and systems, application norms and ethical guidelines.” This may suggest a push to fast-track domestic AI safety rules. Still, given that standard-setting agencies already have detailed workplans, it remains unclear whether this signaling will materially speed up implementation.

Xi’s speech leaves open what specific AI risks are at issue. He names no particular technologies—such as LLMs—and offers no concrete risk scenarios. While current regulations focus mainly on political content, privacy, and bias, Chinese standard-setting bodies and national security experts increasingly reference advanced risks, including existential threats. It is thus plausible—though unconfirmed—that frontier risks were part of this session’s agenda as well.

Overall, Xi’s speech amplifies rather than redirects China’s AI safety posture, reiterating existing priorities under a stronger spotlight.

KEVIN NEVILLE
Writer and Researcher

China’s AI development is often framed as an exceptional or uniquely concerning phenomenon. This risks overlooking the broader forces shaping these policy decisions.

China’s investments in AI, data infrastructure, and automation are best understood as part of a global pattern: Governments and industries worldwide are responding to intensifying economic volatility, shifting geopolitical alignments, environmental stresses, and evolving technology landscapes. These challenges are not unique to China; they are features of a rapidly changing international environment that all major actors must navigate.

The Politburo’s second AI-focused study session reflects this dynamic. Echoing goals set out in the 2017 New Generation AI Development Plan—innovation leadership, ethical governance, and international cooperation—Xi Jinping’s recent remarks signal sharpened urgency. References to "grasping the initiative," "controllability," and "self-reliance" reflect pressures mounting across the system: US-led semiconductor controls, the need for breakthroughs in foundational software and hardware, and growing awareness (if not expectation) of techno-fragmentation.

China’s framing of AI as an "international public good" fits broader diplomatic efforts, but the intensified focus on internal capacity building reveals deeper structural concerns. This is not merely about competition; it is about managing exposure to systemic risks—geopolitical, technological, and environmental—that no major actor can fully escape.

In this light, China's AI strategy mirrors broader international efforts to secure resilience in an increasingly unstable world. More durable insights may arise from understanding these dynamics beyond a narrow lens of anomalous rivalry.

EDITOR'S NOTES

[1] 新型举国体制 – Translated by CSET as "new structure for leveraging national capabilities" and elsewhere as "new whole state system." See e.g. Zhang, L., & Lan, T. (2022). The new whole state system: Reinventing the Chinese state to promote innovation. Environment and Planning A, 55(1), 201-221. https://doi.org/10.1177/0308518X221088294

[2] DigiChina generally translates the 自主 in 自主可控 as independent or indigenous. The clear meaning is independent of problematic foreign dependencies or interference.

[2] 安全、可靠、可控。– The word “safe” 安全 here can also be translated as “secure”. In 2018, DigiChina translated the phrase as “secure, reliable, and controllable.” The appropriate English translation of the term is challenging, rarely perfect, and varies by context.

Subscribe now

Share

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu.

As the global AI safety community and world leaders prepare to convene at the AI Action Summit in Paris on February 10–11, the question of who will represent China has resurfaced. Over the past 15 months, a growing number of national AI Safety Institutes (AISIs) from around the world have worked together to promote technical research and evaluations of advanced machine learning models. Yet as these efforts morphed into a formalized network of national AISIs, China’s involvement in related convenings and joint testing exercises remained uncertain.

Ahead of the Paris AI Action Summit, the Chinese government appears to have finally put forth its equivalent of an AISI. The China AI Safety and Development Association (中国人工智能发展与安全研究网络, CNAISDA), which is online at cnaisi.cn, describes itself as “representing China in dialogue and collaboration with AI security research institutions around the world.”1 On an official list of Paris summit side events and an event registration page, it is labeled as “the Chinese equivalent of the AI Safety Institute.”

Details are scarce for now: The group’s website is only in English and contains limited information. We could find only a few references to the organization on Chinese social media, where its Chinese-language name translates to the “China AI Development and Safety Research Network.” (See below for more on the name and the translation of “safety.”) It is unclear who, if anyone, leads the group. But CNAISDA appears to be China’s answer to the AISI.

More details are likely to surface at its inaugural event on February 11, but here’s what we know so far about this unusual institution. …

[Read the full essay here.]

Subscribe now

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu.

In “Moving Data, Moving Target,” Samm Sacks, Krystal Chen Zeng, and Graham Webster examine a turnabout in the past year in China’s regulation of outbound data flows. Long story short: The Chinese government appeared to recognize it had erected unworkable barriers to business and “eased restrictions affecting many businesses, while still underscoring the strength of the [Cyberspace Administration of China]’s authority over high-risk areas,” we write.

Still, the basic structure of the laws underlying revised 2024 Provisions, which replaced more stringent Measures from 2022, leaves several active questions, each applying to at least some firms doing business across China’s border:

• Eight years after the Cybersecurity Law was finalized, the consequential category of “important data” can still be elusive.

• Those operating with more than 1 million individuals’ worth of personal information are still subject to an uncertain security review system before data can be transferred abroad.

• Free trade zones within China promise the potential for streamlined rules, but their extent and how to qualify for them remains a question.

• The law still leaves the government room to determine whether a transfer is “necessary,” and there’s no clear standard for this.

All of these uncertainties are most likely to be illuminated through the actual practice of regulators and regulated entities. For now, we offer a snapshot in an evolving regime.

Read our full essay here.

Please don’t hesitate to share comments and feedback.

–Graham Webster, Editor-in-Chief, DigiChina

Subscribe now

Share

About DigiChina

Housed within the Program on Geopolitics, Technology, and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu. 

Today, we are releasing a follow-up that looks at subsequent developments specifically in the area of data governance and cross-border data transfers. Samm Sacks, Yan Luo, and Graham Webster map the most impactful measures by both governments to erect barriers or enforce safeguards when it comes to transferring data between the two countries.

This report comes a day after the Biden administration released a major executive order and advance notice of proposed rulemaking that stands to curtail data broker sales to China and other countries of concern, among other measures. This action breaks from longstanding U.S. patterns of regulation, in which little or no data restrictions were focused on specific countries. This report is thus a snapshot in time, but it is the first effort we are aware of to tally up obstacles to U.S.-China cross-border data flows in one place.

You can read the full report in PDF here.

Share

Subscribe now

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu. 

Subscribe now

This DigiChina Forum includes contributions from Jason Zhou, Mingli Shi, Hunter Dorwart, and Johanna Costigan. New contributions may be forthcoming at the link above. It references a guest translation of the draft at hand, available here.

A few months after the introduction of OpenAI's ChatGPT captured imaginations around the world, China's State Council quietly announced that it would work toward drafting an Artificial Intelligence Law. The government had already acted relatively quickly, drafting, significantly revising, and finally implementing on August 15 rules on generative AI that build on existing laws. Still, broader questions about AI's role in society remain, and the May announcement signaled that more holistic legislative thinking was on the horizon.

A team from the Chinese Academy of Social Sciences (CASS) this month released a scholars' draft of an AI Law for China. When the Chinese government announces that it will draft a law, the future of the effort is uncertain. In some cases, one or more groups of scholars drafts up a proposal. These sometimes feed directly into legislative work and their influence is seen in an official National People's Congress draft for public comment. Sometimes—as, for example, with an early 2000s effort toward a Personal Information Protection Law (PIPL)—the process falters. In the case of the PIPL, CASS scholar Zhou Hanhua described in a DigiChina interview how a team was asked by a government office to work on the issue, producing a draft in 2005, before the PIPL drafting effort stalled for years. By the late '10s when the effort was picked up again, technology and law had changed so much that their 2005 draft was not fit for purpose, and little of its content is visible in the law that went into effect in 2021.

In the case of this scholars' draft of an AI Law, the accompanying explanation notes that it is to serve as a reference for legislative work and is expected to be revised in a 2.0 version. Although the connection between this text and any eventual Chinese AI Law is uncertain, its publication from a team led by Zhou Hui, deputy director of the CASS Cyber and Information Law Research Office and chair of a research project on AI ethics and regulation, makes it an early indication of how some influential policy thinkers are approaching the State Council-announced AI Law effort.

We invited DigiChina community members to share their analysis of the scholars' draft, a translation of which was led by Concordia AI and is published here. Their responses are below. –Graham Webster

JASON ZHOU
Senior Research Analyst, Concordia AI

This draft model law is the first step of a likely lengthy and complicated process for China to develop an Artificial Intelligence Law. Given the early stage (and non-governmental nature) of this draft, it is perhaps most useful to focus on the law’s structure and the key risks it tackles. Overall, the draft model law attempts a flexible and pragmatic approach, while also including provisions to mitigate frontier risks from increasingly advanced AI models. 

The key regulatory tool in the draft model law is an AI research, development, and provision “negative list.” Items on the list would require a government-approved permit before being developed or released. As the drafters note in their explanation, this mechanism seeks to balance between safety and development, while keeping safety as the bottom line. Therefore, the drafters adopt a differentiated approach based on risk, which bears some resemblance to the approach taken in the EU AI Act. Whether the model law strikes the right balance between fostering innovation and ensuring safety will depend on the exact composition of the negative list—a challenging task for regulators. 

In that case, what AI capabilities are considered more high risk? The draft model law is hazy on this point, declining to specify how the negative list would be developed and its relationship to the Ministry of Science and Technology’s S&T ethics reviews. From the principles listed in Chapter 1 of the model law, it is clear that adhering to socialist core values, privacy, anti-discrimination, transparency, data security, and human control are all major concerns. Meanwhile, other provisions focus on frontier AI capabilities. Article 43 sets out more stringent oversight of “foundation models,” which though not defined in the draft, are understood to be models trained on large amounts of data that can be used for many different applications and are the subject of a United Kingdom expert task force. Foundation models would require a yearly “social responsibility report” by an independent body of primarily external members. Articles 25 and 50 also call for ensuring human control of AI through both human oversight and technical means, including when AI operates autonomously. As my colleagues at Concordia AI are documenting in forthcoming work, several influential Chinese experts have already expressed concerns about frontier AI risks, and this model law’s provisions on foundation models and human control offers further evidence of a degree of convergence between expert Chinese and international views. However, technical and policy implementation of such safeguards remains to be seen.

China’s existing system of an algorithm registry and security reviews for certain generative AI products already possesses strong policy levers, which a negative list system proposed by the model law could augment if effectively implemented. However, no matter how complete the legal framework, the rapid pace of AI development demands agility, and China will likely adjust regulations and execution as different concerns are prioritized. Thus, understanding Chinese perceptions of AI risks is a critical complement to analyzing laws and regulation, and foundation models are especially critical given their outsized role in AI development. 

MINGLI SHI

The draft model law suggests the formation of a new agency, the National AI Office, to coordinate and supervise the administration of AI technology. The aim is to prevent the potential regulatory chaos, often likened to the mythical notion of “九龙治水” (jiǔlóngzhìshuǐ, nine dragons governing water), where multiple regulatory bodies assert administrative authority over the same subject. A similar concern was taken into consideration during the development of China’s privacy legislation, with respected scholars like Professor Zhou Hanhua advocating for a new agency focused on personal data protection.

Despite this awareness and later endeavors, as well as the fact that the Personal Information Protection Law authorizes the Cyberspace Administration of China to coordinate relevant regulatory work, the supervision of China’s privacy landscape remains fragmented across various government bodies including the State Administration for Market Regulation, the Ministry of Industry and Information Technology, the Ministry of Public Security, and other sector-specific regulators. Therefore, to achieve the purpose of the model law drafters, the new authority, aka the National AI Office would need to be explicitly granted concrete powers of investigation, regulation, and enforcement. At the same time, it is also crucial to limit the authority of existing government bodies that might have skin in the AI game. Without these measures, the nine-dragons-governing-water problem would persist, and the new agency might also struggle to effectively facilitate inter-agency collaboration and coordination.

Additionally, the draft model law introduces the idea of a “negative list” approach, to prevent the regulations from stifling the progress of AI technology. Activities on the list would require specific administrative approval, while those not included would only need to undergo regulatory filing. The success of this approach clearly hinges on what the list would look like, which could tip the balance in either direction. If the list becomes extensive, broad, or vague—causing confusion or uncertainties for affected entities—it might still hold back technological progress. Right now, the draft model law lacks details or criteria on this question.

HUNTER DORWART
Associate, Bird & Bird

This expert draft AI Law provides an important window into the approach Chinese policymakers may take with respect to future AI rules and the extent to which these rules converge and diverge with other global regulatory approaches. While much can be said about the scope, application, and requirements of the draft law itself, one important question is how the separate legal framework for AI will relate to existing regulations (including those that cover similar digital technologies) both within and outside of China.

On the one hand, the expert draft law defers in many places to technical standards and regulations already in force—indicating that organizations will be expected to implement these where appropriate. These instruments generally apply to a narrower group of services or technical components. For instance, Chinese authorities have recently issued guidance on machine learning, algorithm ethics, and labeling and annotating training data, while promulgating regulations on recommender systems, deep synthesis, and generative AI. With the draft AI law, this sector-specific approach would be complemented by a risk-based system for AI products. Although there is no list of prohibited applications, some services will need to obtain a license and submit to heightened screening before deploying to market.

On the other hand, the expert draft law introduces provisions that overlap with Chinese data protection law, which could raise new challenges for organizations. Some of these obligations can be folded into pre-existing compliance programs. Disclosing transparency information in privacy policies; conducting multi-purpose impact assessments; maintaining processing records and training logs; appointing personnel responsible for data protection and AI ethics; implementing security measures; and providing information to regulatory authorities all come to mind. Additionally, the principles-based approach set forth in the expert draft law resonates with existing standards. Covered entities—including AI researchers, developers, providers, and users (a typology also found in the EU AI Act)—must generally abide by varying requirements, such as explaining how the AI products work and ensuring fairness in input data for algorithm training.

However, there are other areas where the provisions will be challenging for organizations to reconcile with their existing compliance obligations. Allocating responsibilities with entrusted processors under the PIPL and the different roles in the expert draft law may raise novel questions for apportioning liability in commercial agreements, while the potential overlap of different filing requirements and security assessments will make global compliance programs more difficult to operationalize. As an example, one of the stipulated goals of AI governance in China is to put safeguards on the use of automated decision-making (ADM) that negatively impacts individuals. Already the PIPL contains provisions to address this—but to my knowledge, these have not been enforced.

Overall, it remains to be seen how Chinese policymakers will fashion AI rules to balance between the goals of technological development, national security, and the protection of rights and interests of individuals, but in a way that complements China’s larger framework for data governance. It will likewise be critical to gauge how enforcement oversight will be divided and co-managed by different bodies, including the newly formed Data Bureau and the Cyberspace Administration of China (CAC). We can expect greater clarity in the near future.

JOHANNA COSTIGAN
Writer and editor

Like China’s actual AI regulations, the “Model Law” from a team at the Chinese Academy of Social Sciences (CASS) shows that Chinese AI researchers share unsurprisingly parallel concerns with their international counterparts. Everyone wants to innovate (and profit-generate) while maintaining safety (or appearing to). 

The CASS authors propose toeing this line by creating an adjustable “negative list” of AI firms and services that are off-limits, thereby clarifying that everything else is not. Wang Jun, one of the authors, noted that the AI industry’s scalability is currently limited by consumers’ distrust in it. Without being surreptitious, he linked transparent regulations with industrial development. “Clear and specified obligations can significantly improve the business environment,” he said. 

That sentiment stands in stark contrast to the “pause” and “risk mitigation” letters that western AI business leaders, among others, have endorsed. Signed by Elon Musk and Sam Altman, respectively, these letters treat AI like it’s a faraway meteor heading to earth instead of a product companies have actively developed and deployed. 

As researchers like Emily Bender have pointed out, the problem with these letters lies in their support of using “longtermist” approaches to control AI. That logic enables an erasure of known, current problems and quandaries introduced by artificial intelligence in service of humoring the sometimes boyish, other times valid visions of worst-case scenarios. Even as it tries to both stimulate and regulate AI, the CASS Model Law usefully lacks comparable hypocrisy—probably because it is a fully-conceived draft law written by researchers rather than a one-lined doomsday prediction. 

The Model Law proposes that China establish a National Artificial Intelligence Office (a worldwide first), stipulates separate requirements for developers and providers, and creates a negative list. These changes would make AI development more transparent—if not to the people, then at least to the party-state. 

The amount of good that could do is debatable. But it is surely better than the United States’ zero-regulation approach, which has helped dangerous narratives with conspicuous industry support gain traction. These range from Altman’s warning that regulation might “slow down American industry in such a way that China or somebody else makes faster progress” to pointing at a possible apocalypse to distract from reality. It will be worth watching how many of the CASS proposals end up in the official draft AI law once it is introduced to the NPC. Hopefully, the final version maintains this draft’s emphasis on transparency—and immediacy. 

Either way, though it is by no means unfiltered, the CASS draft shows that at least parts of China’s AI community are turning over the right stones. That’s also the case in the U.S. The difference is that the Americans making good points do not yet have lawmakers’ full attention. Maybe China’s eventual AI law will change that by inspiring a regulatory layer to the U.S.-China AI competition. For political-logistical reasons, any lessons learned from China’s regulatory experiments would have to remain unsaid.

Subscribe now

Share

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at Stanford University, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu. 

Subscribe now

This translation is by (in randomized order), Seton Huang, Helen Toner, Zac Haluza, and Rogier Creemers, and was edited by Graham Webster. During editing, an alternative translation from China Law Translate was consulted.

Translation

Measures for the Management of Generative Artificial Intelligence Services (Draft for Comment)

Article 1: In order to stimulate the healthy development and standardized application of generative artificial intelligence (AI), on the basis of the Cybersecurity Law of the People’s Republic of China, the Data Security Law of the People’s Republic of China, the Personal Information Protection Law of the People’s Republic of China, and other such laws and administrative regulations, these Measures are formulated.

Article 2: These Measures apply to the research, development, and use of products with generative AI functions, and to the provision of services to the public within the [mainland] territory of the People’s Republic of China.

Generative AI, as mentioned in these Measures, refers to technologies generating text, image, audio, video, code, or other such content based on algorithms, models, or rules.

Article 3: The State supports indigenous innovation, broad application, and international cooperation in foundational technologies such as AI algorithms and frameworks, and encourages the prioritized use of secure and reliable software, tools, computing, and data resources.

Article 4: The provision of generative AI products or services shall abide by the requirements of laws and regulations, respect social virtue and good public customs, and conform to the following requirements:

1. Content generated through the use of generative AI shall reflect the Socialist Core Values, and may not contain: subversion of state power; overturning of the socialist system; incitement of separatism; harm to national unity; propagation of terrorism or extremism; propagation of ethnic hatred or ethnic discrimination; violent, obscene, or sexual information; false information; as well as content that may upset economic order or social order.

2. In processes such as algorithm design, selecting training data, model generation and model optimization, service provision, etc., adopt measures to prevent the emergence of discrimination on the basis of race, ethnicity, religious belief, nationality, region, sex, age, or profession.

3. Respect intellectual property rights and commercial ethics; advantages in algorithms, data, platforms, etc., may not be used to engage in unfair competition. 

4. Content generated through the use of generative AI shall be true and accurate, and measures are to be adopted to prevent the generation of false information. 

5. Respect the lawful rights and interests of others; prevent harm to the physical and mental health of others, infringement of their likeness rights, reputation rights and personal privacy, as well as infringement of intellectual property rights. It is prohibited to illegally obtain, divulge or use personal information and private [information], as well as commercial secrets.

Article 5: Organizations or individuals that use generative AI to provide services such as chat, text, image, or audio generation (hereinafter referred to as “providers”); including providing programmable interfaces [i.e., APIs] and other means which support others to themselves generate text, images, audio, etc.; bear responsibility as the producer of the content generated by the product. Where personal information is involved, they bear legal responsibility as personal information handlers and are to fulfill personal information protection obligations.

Article 6: Before using generative AI products to provide services to the public, a security assessment must be submitted to the state cyberspace and information department [i.e., the Cyberspace Administration of China] in accordance with the Provisions on the Security Assessment of Internet Information Services With Public Opinion Properties or Social Mobilization Capacity, and the procedures of algorithm filing, modification, and cancellation of filing must be carried out in accordance with the Internet Information Service Algorithmic Recommendation Management Provisions.

Article 7: Providers shall bear responsibility for the legality of the sources of generative AI product pre-training data and optimization training data.

Data used for generative AI product pre-training and optimization training shall satisfy the following requirements:

1. Conforming to the requirements of the Cybersecurity Law of the People’s Republic of China and other such laws and regulations;

2. Not containing content infringing intellectual property rights;

3. Where data includes personal information, the consent of the personal information subject shall be obtained, or other procedures conforming with the provisions of laws and administrative regulations followed;

4. Be able to ensure the data’s veracity, accuracy, objectivity, and diversity;

5. Other supervision requirements of the state cybersecurity and informatization department concerning generative AI functions and services.

Article 8: When human annotation is used in the development of generative AI products, providers shall formulate clear, specific, and practicable annotation rules conforming to the requirements of these Measures; necessary training of annotation personnel shall be conducted; and the validity of annotation content shall be spot checked.

Article 9: When providing generative AI services, users shall be required to provide real identity information in accordance with the provisions of the Cybersecurity Law of the People’s Republic of China.

Article 10: Providers shall explicitly disclose the user groups, occasions, and uses for their services, and adopt appropriate measures to prevent users from excessive reliance on or addiction to generated content.

Article 11: In the process of providing services, providers have the duty to protect information input by users and usage records. They may not illegally preserve input information from which it is possible to deduce the identity of users, they may not conduct profiling on the basis of information input by users and their usage details, and they may not provide information input by users to others. Where laws or regulations provide otherwise, those provisions are to be followed.

Article 12: Providers may not engage in content generation that is discriminatory based on a user’s race, nationality, sex, etc.

Article 13: Providers shall establish mechanisms for receiving and handling user complaints and promptly handle individual requests concerning revision, deletion, or masking of their personal information; and when they discover or learn that generated text, images, audio, video, etc., infringe other persons’ likeness rights, reputation rights, personal privacy, or commercial secrets, or do not conform to the demands of these Measures, they shall adopt measures to cease generation and prevent the expansion of the harm. 

Article 14: Providers shall, throughout the lifecycle, provide secure, stable and sustained services, and ensure users’ normal usage.

Article 15: When generated content that does not conform to the requirements of these Measures is discovered during operations or reported by users, aside from adopting content filtering and other such measures, repeat generation is to be prevented through such methods as optimization training within three months.

Article 16: Providers shall mark generated images, videos, and other content in accordance with the Internet Information Service Deep Synthesis Management Provisions.

Article 17: Providers shall, in accordance with the requirements of the state cybersecurity and informatization department and relevant responsible departments, provide necessary information that could influence users trust or choices, including descriptions of the source, scale, type, quality, etc., of pre-training and optimization training data; rules for human annotation; the scale and type of human-annotated data; and foundational algorithms and technological systems. 

Article 18: Providers shall guide users to scientifically understand and rationally use content generated by generative AI; not to use generated content to damage others’ image, reputation, or other lawful rights and interests; and not to engage in commercial hype or improper marketing.

When users discover generated content that does not meet the requirements of these measures, they have the right to report this to cybersecurity and informatization departments or relevant responsible departments.

Article 19: If a provider finds that a user has used generative AI products to violate laws or regulations; violate business ethics or social virtue, including engaging in online hype, malicious posting and commenting, creating spam, or writing malicious software; or engage in improper business marketing; etc.; service shall be suspended or terminated.

Article 20: If a provider violates the provisions of the Measures, the cybersecurity and informatization department and relevant responsible departments are to impose penalties in accordance with the provisions of Cybersecurity Law of the People’s Republic of China, the Data Security Law of the People’s Republic of China, the Personal Information Protection Law of the People’s Republic of China, and other such laws and administrative regulations.

Where there are no provisions of law or administrative regulation, the cybersecurity and informatization department and relevant responsible departments are to, in accordance with their duties, issue warnings, circulate criticisms, and order corrections within a set period of time. Where corrections are refused or circumstances are grave, they are to order suspension or termination of their use of generative AI provider services, and a penalty more than 10,000 yuan and less than 100,000 yuan is to be imposed. Where behavior constitutes a violation of public security management, public security management penalties are to be imposed in accordance with the law. Where a crime is constituted, criminal responsibility shall be pursued in accordance with the law.

Article 21: These measures are effective beginning [day] [month], 2023.

Chinese-language Original

[Source] [Archived copy]

生成式人工智能服务管理办法（征求意见稿）

第一条 为促进生成式人工智能健康发展和规范应用，根据《中华人民共和国网络安全法》《中华人民共和国数据安全法》《中华人民共和国个人信息保护法》等法律、行政法规，制定本办法。

第二条 研发、利用生成式人工智能产品，面向中华人民共和国境内公众提供服务的，适用本办法。

本办法所称生成式人工智能，是指基于算法、模型、规则生成文本、图片、声音、视频、代码等内容的技术。

第三条 国家支持人工智能算法、框架等基础技术的自主创新、推广应用、国际合作，鼓励优先采用安全可信的软件、工具、计算和数据资源。

第四条 提供生成式人工智能产品或服务应当遵守法律法规的要求，尊重社会公德、公序良俗，符合以下要求：

（一）利用生成式人工智能生成的内容应当体现社会主义核心价值观，不得含有颠覆国家政权、推翻社会主义制度，煽动分裂国家、破坏国家统一，宣扬恐怖主义、极端主义，宣扬民族仇恨、民族歧视，暴力、淫秽色情信息，虚假信息，以及可能扰乱经济秩序和社会秩序的内容。

（二）在算法设计、训练数据选择、模型生成和优化、提供服务等过程中，采取措施防止出现种族、民族、信仰、国别、地域、性别、年龄、职业等歧视。

（三）尊重知识产权、商业道德，不得利用算法、数据、平台等优势实施不公平竞争。

（四）利用生成式人工智能生成的内容应当真实准确，采取措施防止生成虚假信息。

（五）尊重他人合法利益，防止伤害他人身心健康，损害肖像权、名誉权和个人隐私，侵犯知识产权。禁止非法获取、披露、利用个人信息和隐私、商业秘密。

第五条 利用生成式人工智能产品提供聊天和文本、图像、声音生成等服务的组织和个人（以下称“提供者”），包括通过提供可编程接口等方式支持他人自行生成文本、图像、声音等，承担该产品生成内容生产者的责任；涉及个人信息的，承担个人信息处理者的法定责任，履行个人信息保护义务。

第六条 利用生成式人工智能产品向公众提供服务前，应当按照《具有舆论属性或社会动员能力的互联网信息服务安全评估规定》向国家网信部门申报安全评估，并按照《互联网信息服务算法推荐管理规定》履行算法备案和变更、注销备案手续。

第七条 提供者应当对生成式人工智能产品的预训练数据、优化训练数据来源的合法性负责。

用于生成式人工智能产品的预训练、优化训练数据，应满足以下要求：

（一）符合《中华人民共和国网络安全法》等法律法规的要求；

（二）不含有侵犯知识产权的内容；

（三）数据包含个人信息的，应当征得个人信息主体同意或者符合法律、行政法规规定的其他情形；

（四）能够保证数据的真实性、准确性、客观性、多样性；

（五）国家网信部门关于生成式人工智能服务的其他监管要求。

第八条 生成式人工智能产品研制中采用人工标注时，提供者应当制定符合本办法要求，清晰、具体、可操作的标注规则，对标注人员进行必要培训，抽样核验标注内容的正确性。

第九条 提供生成式人工智能服务应当按照《中华人民共和国网络安全法》规定，要求用户提供真实身份信息。

第十条 提供者应当明确并公开其服务的适用人群、场合、用途，采取适当措施防范用户过分依赖或沉迷生成内容。

第十一条 提供者在提供服务过程中，对用户的输入信息和使用记录承担保护义务。不得非法留存能够推断出用户身份的输入信息，不得根据用户输入信息和使用情况进行画像，不得向他人提供用户输入信息。法律法规另有规定的，从其规定。

第十二条 提供者不得根据用户的种族、国别、性别等进行带有歧视性的内容生成。

第十三条 提供者应当建立用户投诉接收处理机制，及时处置个人关于更正、删除、屏蔽其个人信息的请求；发现、知悉生成的文本、图片、声音、视频等侵害他人肖像权、名誉权、个人隐私、商业秘密，或者不符合本办法要求时，应当采取措施，停止生成，防止危害持续。

第十四条 提供者应当在生命周期内，提供安全、稳健、持续的服务，保障用户正常使用。

第十五条 对于运行中发现、用户举报的不符合本办法要求的生成内容，除采取内容过滤等措施外，应在3个月内通过模型优化训练等方式防止再次生成。

第十六条 提供者应当按照《互联网信息服务深度合成管理规定》对生成的图片、视频等内容进行标识。

第十七条 提供者应当根据国家网信部门和有关主管部门的要求，提供可以影响用户信任、选择的必要信息，包括预训练和优化训练数据的来源、规模、类型、质量等描述，人工标注规则，人工标注数据的规模和类型，基础算法和技术体系等。

第十八条 提供者应当指导用户科学认识和理性使用生成式人工智能生成的内容，不利用生成内容损害他人形象、名誉以及其他合法权益，不进行商业炒作、不正当营销。

用户发现生成内容不符合本办法要求时，有权向网信部门或者有关主管部门举报。

第十九条 提供者发现用户利用生成式人工智能产品过程中违反法律法规，违背商业道德、社会公德行为时，包括从事网络炒作、恶意发帖跟评、制造垃圾邮件、编写恶意软件，实施不正当的商业营销等，应当暂停或者终止服务。

第二十条 提供者违反本办法规定的，由网信部门和有关主管部门按照《中华人民共和国网络安全法》《中华人民共和国数据安全法》《中华人民共和国个人信息保护法》等法律、行政法规的规定予以处罚。

法律、行政法规没有规定的，由网信部门和有关主管部门依据职责给予警告、通报批评，责令限期改正；拒不改正或者情节严重的，责令暂停或者终止其利用生成式人工智能提供服务，并处一万元以上十万元以下罚款。构成违反治安管理行为的，依法给予治安管理处罚；构成犯罪的，依法追究刑事责任。

第二十一条 本办法自2023年 月 日起实施。

About DigiChina

Housed within the Program on Geopolitics, Technology and Governance at the Stanford University Cyber Policy Center, DigiChina is a cross-organization, collaborative project to provide translations, factual context, and analysis on Chinese technology policy. More at digichina.stanford.edu.

Subscribe now

How to Rebuild China’s ‘Innovation System’–According to Officials and Scholars

Chinese experts add context to this year’s reconfiguration of the country’s science and technology bureaucracy

By Zac Haluza

Editor's note: This is the first in a series of DigiChina contributions by Zac Haluza, newly serving as an Associate Editor of DigiChina. Haluza's independent analysis on Chinese technology policy is published on his newsletter, formerly known as Cloudology.

Since the 20th Party Congress last October, innovation in science and technology (S&T) has become a core theme in Chinese government policy and messaging. In his report during the Congress, General Secretary Xi Jinping laid out the country’s priorities, saying S&T must remain China’s top productive force, talent must remain its top resource, and innovation must remain its top motivator. Another key message from Xi: China needs to achieve “a high level of self-reliance in science and technology.” This same messaging was repeated throughout the Two Sessions in February and March, and these sentiments have reverberated through state agencies and local governments ever since.

As an engine to drive this innovation, Xi has pointed to the importance of China’s “innovation system,” a key term in Chinese policy discourse since at least the 2006-2020 plan for S&T growth. When the CCP Central Committee and the State Council announced their state restructuring plan during the Two Sessions, it included a comprehensive reorganization of the Ministry of Science and Technology (MOST) and established the CCP Central Commission for Science and Technology, a high-level body charged with strategy and reform, as well as with the overall planning of China’s national innovation system. This bureaucratic rearrangement signals a greater emphasis and sharper focus on S&T and a new path for related industries in China.

In addition to these top-level announcements, some of the early details about what the changes to the system mean come from writings by MOST officials in prominent outlets. 

A narrative of China’s science and technology history—from the bureaucracy charged with its future

On March 17, the official bulletin of the Chinese Academy of Sciences published an article (also available here and archived here) discussing the innovation system’s history, structure, and objectives. The authors include He Defang, a MOST deputy secretary general, as well as two other ministry officials and scholars from the Shanghai Institute for Science of Science and the National Center for Science and Technology Evaluation.

SOURCE:
"Analysis and Thinking on Developing Evolution of National Innovation System 国家创新体系的发展演进分析与若干思考"
By He Defang 贺德方, Tang Fuqiang 汤富强, Chen Tao 陈涛, Luo Xianfeng 罗仙凤, and Yang Fangjuan 杨芳娟
Bulletin of the Chinese Academy of Sciences 中国科学院院刊, 2023 38(2): 241-254, doi: 10.16418/ j.issn.1000-3045.20230113002. [archived]

The article is a timely outlay of how Chinese government experts view the country’s S&T progress, the challenges it now faces in these fields, and how to best tackle them.

Early in the piece, the authors outline the development of China’s innovation system. A key aspect of this development, they write, is the evolution of the entities that drive China’s innovation—a group that now includes enterprises, research institutes, and institutions of higher education.

From the nation’s founding in 1949 until Deng Xiaoping’s economic reforms, they write, China had a highly centralized administrative system for S&T. Research in these fields was primarily carried out in centralized research institutes; the country’s enterprises were primarily concerned with production. 

With the policies of reform and opening up beginning in the late 1970s, the nation’s industrial and S&T administrative systems experienced major transformations. However, they write, from the beginning of the reforms through the start of the 21st century, no new systems emerged for scientific research and development, as China’s patterns of scientific research still followed the drumbeat of its planned economy. Research institutes, as well as universities and colleges, were the primary entities leading the country’s research and development in S&T fields during this period, they write, but beginning in the late 1980s, China’s policy of market access in exchange for technology (which the U.S. has more recently referred to as “forced technology transfer”) brought a large number of Chinese businesses within what they characterize as a national manufacturing network.

In 2006, the authors write, these dynamics shifted to more closely resemble the landscape we know today. Specifically, China established a “fundamental, systematic framework for innovation in science and technology,” which included a diverse array of agents for innovation: research institutes, enterprises, colleges and universities, as well as new forms of organizations and alliances for research and collaboration. 

Diagnosis and Prescription: Three Layers and Two Loops

Despite this progress, the authors identify persistent problems, which they say this new system intends to address:

“In comparison to the world’s S&T superpowers,1 China’s innovation and development in S&T has no shortage of issues, such as deficiencies in foundational and critical technologies, a lack of interaction between the educational and technical industries, and a shortage of industry members in the community for S&T innovation. These issues have severely restricted the overall effectiveness of the innovation system. In the face of deep and complex changes occurring both within China and abroad, the national innovation system urgently needs to undergo a systematic transformation.”

The authors characterize this revamped innovation system as pragmatic, in which practice takes precedence over theoretical research, and “behavioral agents” such as businesses and universities form and improve their abilities to innovate through their interactions with one another. The purported key to this system’s effective operation lies within the coordinated development of two parallel systems: a “hard” system consisting of S&T ability and a “soft” system made up of laws and policies that spur innovation in science and tech:

“Thus, to effectively transform the innovation system, it is essential to maintain the cooperative evolution of technology and systems, as well as the long-term dynamic coordination between systems and goals of economic development.”

As for the structure of the new framework (diagrammed below), they write that it consists of “three layers and two loops.” The lowest and most foundational layer is that of “environmental support” (in green): market conditions, the protection of property rights, honesty in scientific research, ethics in tech, and so on. This directly supports the second layer, “agents of innovation” (the blue face of the cube), which consists of the innovation-driving institutions discussed above, such as businesses, schools, and national research labs. These agents interact with a global network of innovative entities (the peach block), constituting a “loop” of open innovation. At the same time, these agents’ contributions feed into the third and topmost layer: China’s national objectives and needs (the purple roof). Overseeing and guiding this entire flow is the government (the pink column): through measures such as strategic planning and the guaranteeing of resources, the government engages with all three layers in a “systematic feedback loop.”

A Roadmap for Central Policy Improvements

In the final part of the article, the authors share their thoughts on how China’s innovation system can be strengthened. They highlight the progress that this system has made over the last two decades, while emphasizing the challenges ahead. Bringing to mind the relationship between real technological activity and government policy illustrated above, they emphasize the need to strengthen China’s technological abilities while also improving its policy framework:

“By strengthening legislation for science and technology, creating innovative policy tools, and optimizing the environment for innovation, China can improve mechanisms, policies, and measures that are conducive to innovation, drive the iterative improvement of administration for science and technology, and make the innovation system more efficient.”

Some of their prescriptions for making China’s innovation system more effective include continuing the Party’s centralized leadership and strengthening the planning and development of education, S&T, and talent, as well as boosting enterprises’ competitiveness and innovative capabilities. For those familiar with MOST rhetoric on China’s S&T development, these solutions will certainly not come as a surprise.

Experts on the Same Page

Around the same time that the Chinese Academy of the Sciences published this article in their bulletin, they also collaborated with the state-run China Internet Information Center to interview Fan Jianping 樊建平, an expert in the fields of high-performance computing, cloud, and distributed computing who leads the Academy’s Shenzhen Institute of Advanced Technology.

SOURCE:
"[Hongyi] Fan Jianping: The Innovation System Should Have an Active Government and Effective Market Fulfill Their Function【闳议】樊建平：创新体系建设应发挥有为政府与有效市场的作用"
Text based on the Hongyi video program published on the Bulletin of the Chinese Academy of Sciences WeChat public channel: https://mp.weixin.qq.com/s/80TylLYhTm_lpjLjCNyMiQ [archived]

Fan’s comments on how China should strive to improve its current innovation system echo many of the sentiments described above:

“I believe our national innovation system should express itself in two ways: an active government and an effective market.2 Our nationwide system is more on the government end. That which one should call an innovation system supported by government capital is currently undergoing reconstruction—from the establishment of new national laboratories to each province’s laboratories as well as the reconfiguration of key national labs, to the research and development organizations of enterprises, and even regional research organizations in different cities. I believe that these things have increased the strength of the government’s investments. Things have changed very quickly over these last few years; in this regard, [the national innovation system] should express itself through a facilitating state.

“The second matter concerns the agents of innovation, which are enterprises — in particular, innovative high-tech private enterprises. The matter is how to give play to the innovative abilities of small private enterprises; for instance, creating innovative hotspots, in the same way that Beijing, Shanghai, and Shenzhen have become hotspots for scientific research, technological R&D, and industrialization. When it comes to the market, the aspect of finance is very important. Venture capital funds and the like must also develop. In this way, you have a facilitating state as well as an effective market; in addition, the two things will have the chance to combine with one another and also combine the advantages of China’s national system with the advantages of the market. The next steps should concern how to carry out these two things well and let them properly express themselves. China will likely move faster than the West along the path of innovation.”

Share

Subscribe now

Subscribe now

This article was originally published April 7, 2023, at https://digichina.stanford.edu/work/targeting-u-s-chip-firm-micron-chinas-cybersecurity-reviews-continue-to-evolve/. Please cite the stanford.edu version.

Targeting U.S. Chip Firm Micron, China’s Cybersecurity Reviews Continue to Evolve

Will China's cybersecurity reviews remain granular, or will they be used to ban entire companies?

By Graham Webster

Last week, the Cyberspace Administration of China (CAC) announced a cybersecurity review of the U.S. semiconductor firm Micron and the goods it sells in China. The official announcement manages to reveal little while raising several questions about the nature of the cybersecurity review regime. As a starting point, here is the full announcement in translation:

"Notice on the Initiation of Cybersecurity Review into the Products Sold in China by Micron, Inc.

"March 31, 2023, 8:30 p.m.   Source: Cyberspace Administration of China

"In order to safeguard the supply chain security of critical information infrastructure, to guard against product issues and hidden dangers leading to cybersecurity risks, and to defend national security; in accordance with the National Security Law of the People's Republic of China and the Cybersecurity Law of the People's Republic of China; the Cybersecurity Review Office is implementing a cybersecurity review of the products sold by Micron in China according to the Cybersecurity Review Measures.

"Notice is hereby given."

This is fairly typical language announcing a Chinese regulator’s enforcement action. It gives reasons that echo statutory language, cites legal authorities behind the action at a broad level—and little else.

Looking closer, however, this announcement sets off a new round of questions about the nature of the “cybersecurity review” regime led by the CAC. It underlines continued uncertainty about what might or might not trigger a cybersecurity review. Below, I review the triggers for cybersecurity review, outline the evolution of this regime over time, and analyze how the Micron announcement fits into what we know. Published commentary from a Chinese expert and Chinese media discourse further signals a U.S.–China geopolitical context that cannot be ignored.

What triggers China’s Cybersecurity Review system?

Some triggers for cybersecurity review are clear. The Cybersecurity Review Measures explicitly require critical information infrastructure (CII) operators (i.e. any entity operating IT or data systems across a broad range of key sectors) to file for cybersecurity review when procuring network products and services that may affect a capacious concept of national security. This approach appears to trigger an enormous volume of work, requiring filing for each pairing of an individual product or service with an individual use by a CII operator. If a given piece of network hardware is approved for use in your slightly security-sensitive factory, for example, that doesn’t mean it would be cleared for installation at the heart of the electric grid. This best-developed type of cybersecurity review is fundamentally a granular, risk-based approach to cybersecurity regulation, as opposed to blanket certifications or prohibitions at the product or supplier level. Another clear trigger for review is if a firm handling the personal information of more than 1 million users is headed for a foreign IPO. 

Other triggers for cybersecurity review clearly exist, but they are not specified—and this is not a new uncertainty. The first cybersecurity review to be publicized came in 2021, when the CAC announced a review of DiDi, the Chinese ride-hailing company that had days earlier gone forward with an IPO in New York that regulators had reportedly warned them to delay. Although DiDi clearly handled more than 1 million users’ data going into the IPO process, that explicit trigger for a required filing was added only after the government had slapped DiDi with a review. There was no obvious connection with the preexisting enumerated triggers for review.

Now revised, the current text of the Cybersecurity Review Measures gives no explicit indication that cybersecurity review might be initiated by regulators outside of the contexts where filing is required, but there’s nothing saying it cannot be. Apparently, the Cybersecurity Review Office (CRO), a unit of CAC charged with administering these reviews and consulting with 12 diverse departments, is also able to move based on its own initiative or on other prodding from within the government.

In Micron’s case, based on what little public information is available, several things could have happened:

1. A CII operator might be seeking to procure one or more Micron products for one or more specific uses, filing with the CRO for review as explicitly required. According to the Measures, however, the decision to review in this context would go to the CII operator who would then require the cooperation of Micron (Articles 9 and 6). The recent announcement doesn’t immediately fit this scenario.

2. The CRO may be faced with multiple filings for approval to use Micron products and, separate from the Measures’ described approach of reviewing for each CII operator and each procurement, they might have decided to review the supplier all at once. This would be a deviation from published procedures but still reliant on published triggers for review.

3. Other factors may have led the CRO to decide on a novel form of review, similar to what happened with DiDi, presumably drawing on the vague legal authorities. The published triggers and procedures would not seem to be fully active.

Can't stop moving

A quick timeline of the “national security review” and “cybersecurity review” regime can help explain why something like option 3 is a real possibility, and potentially most likely. Over a period of almost eight years, the business of reviewing technology for national security risks in China has been constantly evolving:

• July 2015 — The National Security Law (NSL) sets out state authorities for “national security review” involving “specified products, critical technologies, and internet information technology products and services” (Article 59).

• November 2016 — The Cybersecurity Law (CSL), which is to take effect the following year, establishes “national security reviews” by the CAC and other departments for “CII operators purchasing network products and services that might impact national security” (Article 35). Details are thin.

• May 2017 — To coincide with the June 1, 2017, effective date of the Cybersecurity Law, the CAC issues the Interim Measures on Security Review of Network Products and Services. These are effective pending unspecified future regulatory work, and they explicitly reference the NSL and CSL.

• May 2019 — The CAC and 11 other regulatory entities issue the Cybersecurity Review Measures (Draft for Comment), which would replace the 2017 Interim Measures if finalized and implemented. DigiChina analyzes the differences between the existing interim measures and the draft at the time.

• April 2020 — The same 12 regulators issue finalized Cybersecurity Review Measures, which replace the Interim Measures on June 1, 2020—three full years after the Cybersecurity Law took effect.

• July 2021 — The Cybersecurity Review Office (CRO) of the CAC, which is established by these Measures, announces a cybersecurity review of DiDi Chuxing two days after the company’s IPO. The Cybersecurity Review Measures do not contain clear and specific grounds for the review.

• July 2021 — Days after it announces the DiDi cybersecurity review, the CAC releases a draft revision to the Cybersecurity Review Measures, which adds the China Securities Regulatory Commission to the dozen regulators involved in the regime and adds the risk of data disclosure to foreign governments through listing on foreign stock exchanges as an explicit grounds for review. It also adds reference to the newly passed Data Security Law.

• December 2021 — The 2021 revision to the Cybersecurity Review Measures is finalized, including the new language on foreign IPOs, and is to take effect Feb. 15, 2022.

• July 2022 — The CAC announces that DiDi’s cybersecurity review resulted in the opening of an investigation resulting in a large fine.It says it found violations of the Cybersecurity Law, the Data Security Law, and the Personal Information Protection Law and announces a fine of 8.026 billion RMB (~$1.2 billion USD).

This progression reveals two important facts. First is that even the most stable period in the cybersecurity review regime’s development (the three years from June 2017 to June 2020) was explicitly uncertain; the Measures in effect then were “interim” in nature and always sat ready to be revised. Second, while these documents vest the CAC’s CRO with responsibility to coordinate reviews and enumerate some specific requirements, the office’s role draws its legal weight from broad provisions of the NSL, CSL, and Data Security Law. While some Chinese regulators may hew more closely to enumerated responsibilities, the DiDi case made clear that at least in some cases, the CRO would adapt to the needs of novel situations. The relevant laws, after all, were drafted in full knowledge that technology and its context would develop rapidly.

Why Micron? Early indications.

Few details are public about the Micron situation, and we can expect to learn more—even if not a lot more—over time. For now, some Chinese commentary can at least suggest hypotheses. Beijing Institute of Technology Law Professor Hong Yanqing, a leading Chinese data governance expert, has written two interesting WeChat posts on the Micron review. In one post, Hong points to the short CAC announcement, which specifies risks to supply chain security for CII, as well problems and hidden dangers in the products, as the area of concern. Article 10 of the Cybersecurity Review Measures lists general types of risks the reviews are concerned with, including several that are relevant here. On supply chain security, there is the risk of “illegal control of, interference with, or destruction of CII,” “the harm to CII business continuity of product and service supply disruptions,” and “the risk of supply disruptions due to political, diplomatic, and trade factors.” In an era when U.S.–China supply chains for semiconductors are acutely affected by U.S. restrictions and may be affected by further limits in the future, supply chain security is not just about sneaking in vulnerabilities.

“Now, as for the reason Micron is being reviewed,” Hong writes, “those following the China–U.S. technological gamesmanship will have their own understanding—specifically, the violent shock to global semiconductor supply chains brought about by the frequent revisions in recent years to the U.S. Export Administration Regulations (e.g. changes to the scope of jurisdiction and territories at any time).” Hong points to an article in the tech news publication ijiwei.com, one of many in Chinese outlets that have discussed the firm’s role in advocating for the unprecedented semiconductor restrictions the U.S. government has implemented on China. The unbylined article argues that Micron has been an active lobbying force for restrictions on China’s tech industries and a direct beneficiary of U.S. policy actions targeting China. It is beyond the scope of the present article to evaluate this narrative or check its assertions, but the prevalence of Chinese commentary identifying Micron as a company particularly unfriendly to China is important in understanding the context for its cybersecurity review.

Finally, Hong’s second commentary focuses on what is likely to happen if Micron does not pass this review: namely, that CII operators would be prohibited from purchasing the company’s products. This raises a question that the Micron case may help answer: Will China’s cybersecurity review regime turn out to operate as it once appeared—a risk-based review system that differentiates between different use cases and different products? Or will it turn out to be a tool that can also effectively ban suppliers in all use cases within the sometimes hard-to-discern boundaries of CII?

If you have not already, please consider subscribing to this newsletter:

Subscribe now

You can also keep track of DigiChina following us on Twitter at @digichn. And if you like what you see, please consider sharing this issue with friends and colleagues:

Share

A new ‘Digital China’ Plan, translated

Ever since people started writing in Chinese about our DigiChina Project, it has been a bit awkward that the usual Chinese-language translation is 《数字中国》(“Digital China”), which also happened to be one of the Chinese government’s tech policy buzzwords beginning around the time of the establishment of the Leading Small Group for Cybersecurity and Informatization in 2014.

Digital China, the concept that shares our modest effort’s name, has in recent years moved to the center of official Chinese discourse on technology. Premier Li Keqiang gave it special emphasis around the 2021 unveiling of the 14th Five-Year Plan, in which it represented one of the 19 overall parts. Now, the Communist Party and the State Council have issued a “Plan for the Overall Layout of Building a Digital China.” The full text of the Plan is not apparently public, but DigiChina has translated the fullest state media report.

DigiChina readers will perhaps unsurprisingly be familiar with much of the Digital China Plan content. Specialists will find a close read rewarding, but two things stand out.

First, this Plan has been issued by the Communist Party of China, as well as the State Council. This stands in contrast to the previous major tech sector plan, the 2017 New Generation Artificial Intelligence Development Plan (AIDP), which was a State Council product, mainly of the Ministry of Science and Technology system. The Party’s role is emphasized in the new Plan:

“Uphold and strengthen the overall leadership of the Party in the building of a Digital China; under the unified and centralized leadership of the Party Central Committee, the Central Commission for Cybersecurity and Informatization [CCCI] is to strengthen overall coordination, comprehensive promotion, and supervision and implementation of building a Digital China.”

Given that the Cyberspace Administration of China is the office of the CCCI and is a one-institution, two-nameplate Party-state hybrid organization—and in light of a WSJ report that China may create a new entity for data governance—we’ll be watching for signs and implications of any shift in Party vs. state roles here. As for other outcomes to track, note that unlike the AIDP, which was released in full and included some fairly concrete metrics against which progress could be measured, much of the Digital China Plan as released is too vague to specifically assess.

Second, in the grand tradition of Chinese planning, we have a new numerical framework: 2522. Two foundations, five integrations, two capacities, and two environments. Study up!

The Plan clarifies that the construction of Digital China will be laid out in accordance with the overall framework of "2522," that is: to use the "two foundations" of digital infrastructure and data resource systems; to promote the deep fusion of the “five integrations” of digital technology with the economy, politics, culture, society, and ecological civilization; to strengthen the "two capacities" of the digital technology innovation system and digital security shield; and to optimize the "two environments" of domestic and international digital development.

(For the record, when naming DigiChina circa 2017, we did not consider that it would likely translate as 数字中国, a term that was already circulating, though much more quietly than “cybersecurity and informatization.” It seems we’ll increasingly have some explaining to do.) –GW

• 📄 DigiChina translation: Plan for the Overall Layout of Building a Digital China

• 🀄️ CAC Deputy Director Cao Shumin writes in Red Banner Manuscripts: “Driving Chinese-style Modernization through Informatization and Digitization” 以信息化数字化驱动引领中国式现代化

• 📋 Pacific Forum has a well-timed paper diving into the “Digital China” concept and arguing it is the world’s first digital grand strategy.

China’s cross-border data system grinds into motion

Since the Cybersecurity Law was finalized in 2016, businesses with any cross-border activity have been waiting to see precisely how the law’s provision governing outbound transfer of certain data (Art. 37) would take shape. Long-awaited measures fleshing out the implementation of this and subsequent cross-border data provisions finally took effect last September, with enforcement for pre-existing data flows taking effect March 1 this year.

Multinational companies reportedly rushed to figure out whether their operations in China were covered and, if so, to complete the filing and review process in time. In January, a cancer research collaboration with a Dutch medical center was the first to have its data transfers approved, Caixin reported, with Air China getting the second approval.

Relatedly, the China Cybersecurity Review Technology and Certification Center was the first institution approved to conduct personal information protection certifications under the Personal Information Protection Law, WilmerHale wrote. And CAC published Measures for Standard Contracts for the Outbound Transfer of Personal Data, as well as a template contract, another long-awaited piece in the cross-border data protection puzzle. –GW

📄 China Law Translate: Translation of the Measures for Standard Contracts for the Outbound Transfer of Personal Data

OpenAI’s ChatGPT is not so open in China

In China, the popularity of ChatGPT has led to substantial public interest in the progress of its own generative artificial intelligence. But if you’re physically in China, it’s a hassle to even poke around with ChatGPT itself: the San Francisco startup OpenAI has banned Chinese (+86) and Hong Kong (+852) numbers for account registrations—meaning although anyone in the PRC can access the OpenAI website that hosts ChatGPT, they can’t use the tool unless they can receive an SMS verification code sent to a non-Chinese phone number. And China isn’t the only country where OpenAI services are restricted: Russia, Vietnam, and Saudi Arabia are not among the 162 supported countries and territories.

Tech-savvy users in China have come up with workarounds, using foreign virtual private networks (VPNs) to bypass IP checks and U.S. VoIP services for SMS verification. They also need to make sure that the frequent switching of VPNs will not trigger OpenAI’s anti-abuse firewall. Some developers have built bots on WeChat and other China-based social media networks using OpenAI’s API, but a handful of them have been suspended by Chinese authorities and their host platforms.

By character count, only some 0.19% of GPT-3’s training dataset is in Chinese, compared to 92% in English. In this context, it’s surprising how well ChatGPT works in Chinese, though still worse than its English-language capabilities. Still, the question remains: Why hasn’t China invented its own ChatGPT? It’s convenient to attribute the Chinese lag in generative AI to its political system and top leader Xi Jinping’s critical view of consumer technology, but that doesn’t explain China’s tremendous progress in AI in recent years. As Louise Matsakis reports for Semafor, regulatory risks, content moderation concerns, U.S. export controls, brain drain, and the priority of profit-driven research are likely contributors to the relative stagnation.

It is, however, very likely China isn't that far behind. Baidu, which has released an image generator and is planning to launch its own chatbot Ernie, might integrate its own version of GPT into Baidu AI Cloud, as Kevin Xu argues in . Other companies including Alibaba, Bytedance, JD, and Qihoo 360 are entering a generative AI arms race alongside Baidu. –TF

What we’re reading

Lessons from EU and Chinese approaches to AI governance. Matt O’Shaughnessy and Matt Sheehan compare AI regulatory regimes—the vertical approach of the AI Act in Europe, which places constraints on each category, and the horizontal approach of China’s algorithm registry.

• 🔍 Carnegie Endowment: Lessons From the World’s Two Experiments in AI Governance

More on EV primacy in China. Chinese authorities are piloting a program that deploys new energy vehicles to public-sector services, including taxis and buses as well as sanitation, postal, and logistics vehicles. 

• 📰 SCMP: China’s plan to fully electrify public vehicles to give NEV sector US$118 billion boost, analyst says

• 📄 MIIT: 装备工业一司相关负责人就《关于组织开展公共领域车辆全面电动化先行区试点工作的通知》答记者问

An Internet of Vehicles. Meanwhile in Jiangsu province, the city of Wuxi has issued guidelines for citywide adoption of “internet of vehicles” technologies—building infrastructure connecting networked cars—that go into effect on March 1. Last February, the MIIT issued guidelines for cybersecurity and data security for the “internet of vehicles,” which, we can be honest, is a better coinage in Chinese. 

• 📰 Sixth Tone: In a First, Chinese City Unveils Law for Internet of Vehicles

• 📄 National People’s Congress: 《无锡市车联网发展促进条例》3月1日施行

• 📄 Wuxi People’s Government: 无锡市车联网发展促进条例

• 🔍 Grid: How China built the world’s biggest EV charging network - and left the US far behind 

Back in the game. The National Press and Publication Administration (NPPA) approved 87 video game licenses in February—some of them were granted to major players Tencent, NetEase, and Bytedance. It’s among several signs that the video game industry is, at least tentatively, out of the doghouse.

• 📄 NPPA: 2023年2月份国产网络游戏审批信息

• 📰 SCMP: China doles out 87 new video game licences in February, with Tencent and NetEase among the recipients

–TF

Subscribe now

Share

If you have not already, please consider subscribing to this newsletter:

Subscribe now

You can also keep track of DigiChina by following us on Twitter at @digichn. And if you like what you see, please consider sharing this issue with friends and colleagues:

Share

DigiChina Forum: Is the rectification over?

For the platform economy, data-intensive businesses, and key consumer-facing sectors such as fintech, the last two years have been turbulent in China. A series of major regulatory moves targeting prominent firms such as Ant Group and DiDi coincided with the implementation of new laws on personal information protection and data security. These diverse events have collectively been labeled a “crackdown” or “rectification” in the digital economy.

Since the beginning of 2023, DiDi seems to be out of the regulatory doghouse and is accepting new sign-ups. A top bank regulator said at least part of the rectification is “basically complete.” And although much remains to be seen as to how the new data laws are enforced, a long period when new draft language would open enormous legal uncertainty has yielded to a basically complete scaffolding for cyberspace regulation.

DigiChina welcomed eight specialists to give their answers to the questions:

Is China’s regulatory fervor in tech over, or is it the new normal? What has it wrought, and what’s yet to come?

Read answers from Rogier Creemers of University of Leiden, Johanna Costigan of Asia Society, Paul Triolo of the Center for Strategic and International Studies, Tom Nunlist of Trivium China, Lauren Dudley of Rhodium Group, the threat intelligence analyst Mei Danowski, Martin Chorzempa of the Peterson Institute for International Economics, and Karman Lucero of Yale Law School.

Their views and insights are diverse, but if one thing emerges as a commonly-held view, it’s that while tech fields may see less upheaval, stronger state attention to the digital economy is here to say. –GW

• ⚖️ Read the full forum: Is China’s Tech ‘Crackdown’ or ‘Rectification’ Over?

Looking into reports of a state-backed DiDi competitor

Student Editor Tianyu Fang looked into reports that Xuexi Qiangguo 学习强国, sometimes called the “Little Red App,” may be introducing a ride-hailing service that could compete with DiDi, just after the privately-held market leader got out of app store jail:

“Looking deeper, however, there is little justification for speculation that this amounts to a government-backed alternative to displace DiDi. The Ministry of Transport, whose logo appears on a screenshot of the platform, has denied any involvement; Beijing Daily, the municipal Party Committee–affiliated newspaper that first broke the news, has subsequently removed its coverage. In a Jan. 19 statement, Xuexi Qiangguo said Qiangguo Jiaotong was not a “national-level transport platform,” but a mini-program developed in partnership with the China Traffic Communication Information Center (CTCIC), a body under the Ministry of Transport. The platform wouldn’t have its own services but rather would aggregate third-party providers—and DiDi might even be one of them.” –TF

• 🚕 Read the full analysis: Is a State-Backed Ride-Hailing Service Set to Displace DiDi in China? Think Again.

China is an early mover in regulating media synthesis services

China’s regulations on synthetic media, the Internet Information Service Deep Synthesis Management Provisions, took effect Jan. 10. The provisions govern a wide variety of services that now or in the future give users the ability to algorithmically generate content, ranging from text-generating models similar to OpenAI’s ChatGPT, to face swap apps, deep fake production, or 3D environment generation for virtual or augmented reality.

As I told The Wall Street Journal, these provisions represent Chinese policymakers’ efforts to channel international discussions about how to govern the use of AI into some of the earliest concrete regulations in any major market. The rules represent a more detailed explication of requirements under the Cybersecurity Law, the Data Security Law, the Personal Information Protection Law, and the Internet Information Service Management Measures. At root, they define the types of monitoring and due diligence, as well as incident response, required for entities providing “deep synthesis” media services to users.

Among the most interesting provisions—not at all out of step with regulatory thinking around the world, but notably now in effect in China—is to require prominent labeling of outputs that might confuse or mislead the public (Art. 17), and to prohibit anyone from removing such markers (Art. 18). It will be interesting to watch how this is accomplished if, for instance, reports are true that Baidu will soon release a text-generating product like ChatGPT. –GW

• 📄 DigiChina translated last year’s draft of the Provisions

• 📄 China Law Translate has a version of the final

What we’re following

• Signaling support for SMEs: The Ministry of Industry and Information Technology (MIIT) hosted its annual National Work Conference on Jan. 11, outlining 13 critical tasks in 2023. As noted by Trivium, the conference called for support for small- and medium-sized enterprises, potentially signaling a friendlier regulatory regime for technology entrepreneurs.

  • 📄 MIIT: 全国工业和信息化工作会议在京召开

• 20 articles on data systems: The State Council issued guidelines (containing 20 articles) on “building a basic system to better give play to the utility of data factors [of production]” in December, followed by a document published in Qiushi magazine by the National Development and Reform Commission (NDRC) emphasizing its importance.

  • 📄 State Council: 关于构建数据基础制度更好发挥数据要素作用的意见

  • 📄 NDRC: 加快构建中国特色数据基础制度体系 促进全体人民共享数字经济发展红利

  • 🔍 Analysis from : What Are “Data Systems with Chinese Characteristics”?

• Now you can file an appeal to credit blacklists: New provisional measures on “credit information restoration,” published by the NDRC this month, outline procedures for entities to remove themselves from negative credit information catalogs.

  • 📄 NDRC (via China Law Translate): Measures on the Management of Credit Information Restoration Following the Correction of Untrustworthy Conduct (Provisional)

  • 🔍 China Law Translate analysis: Take it down! Stopping the display of “negative credit information” –TF

Happy 2023 and Year of the Rabbit! Resolve to share and subscribe.

Subscribe now

Share