The Political Limits of China’s AI Diffusion Ambitions

Editor’s Note: Artificial intelligence (AI) is shaking up China’s economy, just as it is raising many questions in the United States about the future of work and the employment market, among other issues. Ruby Scanlon of the Center for a New American Security, however, argues that China’s concerns about employment stability limit rapid AI adoption. Beijing is wrestling with how to use AI for maximum efficiency while avoiding a major increase in the already-high unemployment rate.

Daniel Byman

***

When a Beijing-based company replaced an employee, identified as Liu, with an automated system, it likely expected China’s courts to side with its decision. After all, the government’s AI Plus initiative targets 70 percent AI adoption across key sectors by 2027 and 90 percent by 2030. Instead, the court sided with Liu.

In a 2025 ruling, a Chinese court ruled that replacing a worker with AI is not valid grounds for dismissal and that firms must first attempt contract renegotiation, retraining, or internal reassignment. Beijing’s labor bureau subsequently published the decision as a “model case,” the Chinese judicial system’s rough equivalent to stare decisis in the United States, signaling how similar disputes should be handled. The case is emblematic of a shift in how Chinese policymakers are preparing for the possibility of AI-driven labor displacement and may foreshadow slowed diffusion of AI in China in favor of worker protections and economic stability.

Despite the unemployment risks of AI, the prevailing logic in Beijing frames rapid AI diffusion as essential to sustaining growth as its working-age population shrinks and the economy faces mounting pressure from weak domestic consumption and a prolonged property crisis. But an underappreciated constraint on Beijing’s diffusion agenda is its enduring commitment to employment stability. China’s regime relies on performance legitimacy, a tacit bargain in which citizens accept single-party rule in exchange for rising living standards and economic opportunity. Fear of collective action, such as local protests, labor strikes, and online activism, reinforces this imperative, making Chinese local governments surprisingly responsive to citizen grievances. Field experiments conducted by researchers at UCLA and Hong Kong University found that 36 percent of local queries receive a government response in China, compared to 22 percent in the United States. These two goals, of AI diffusion and social and economic stability, are increasingly in tension as AI threatens to automate millions of jobs.

Early indicators of AI’s impact on China’s labor market suggest reason for concern. Youth unemployment hit a record 18.9 percent in August 2025 following a surge of new graduates; it has since fallen only modestly, to 16.9 percent as of March 2026, remaining nearly double the levels seen decade ago. While multiple factors—from weak private-sector hiring to property-market distress—contribute to these figures, researchers at Stanford University have suggested that early-career workers in AI-exposed occupations may function as “canaries in the coal mine,” offering early signals of broader labor displacement. A 2025 central bank survey likewise showed domestic employment sentiment at a record low, reflecting widespread anxiety about future job prospects, even if causality remains unclear.

The policy landscape may already be shifting in favor of workers. In mid-2024, Baidu’s Apollo Go robotaxi service in Wuhan sparked a social media firestorm after taxi drivers petitioned the city to restrict the service. Authorities subsequently froze the company’s approvals for new autonomous vehicles and its expansion into additional cities for several months in the second half of 2024. On a trip to Beijing in November 2025, I heard from one industry representative that Meituan, China’s largest food delivery platform, has limited the scale of its autonomous delivery vehicle rollouts due to government pressure to protect its millions of human couriers. These cases suggest that even in a system that prioritizes technological advancement, the political costs of visible labor displacement can force meaningful policy retreats.

The Chinese government is increasingly using employment stabilization language alongside statements of its diffusion ambitions to preempt concerns about potential dislocation from AI. Beijing’s AI Plus plan included directives to assess employment risks from AI applications and steer innovation toward job-creating sectors in order to reduce impacts on employment. At the 2025 Two Sessions, an official floated an “AI + Employment” framework that would feature tax incentives, reskilling programs, wage subsidies, and potential limits on automation in certain jobs. Liu Qingfeng, a National People’s Congress representative and founder of tech firm iFlytek, called for “a special AI-unemployment insurance program … [providing] a dedicated fund to protect jobs most vulnerable to AI disruption.” In August 2025, the State Council issued its “Opinions on Deepening the Implementation of the AI+ Action,” which included a directive to “strengthen employment risk assessments for AI applications, guide innovation resources toward directions with greater job-creation potential, and reduce the impact on employment.” In late January 2026, China’s labor ministry announced plans to release a dedicated policy addressing AI’s impact on employment, including measures for “job stabilization, expansion, and quality enhancement.”

In February 2026, Cai Fang, president of the Chinese Association of Labour Economics and arguably China’s most renowned labor economist, weighed in on China’s policy options with increased fervor. He argued that in the not-too-distant future, AI models will possess intelligence equal to or even higher than that of humans, and will increasingly replace all human jobs, requiring Beijing to increase social spending, explore adoption of a universal basic income (UBI), and subsidize “Baumol jobs” (traditionally low-productivity jobs such as performing arts) and industries with “nostalgic value” such as physical book stores. That proposals like UBI and low-productivity job subsidization are now entering mainstream policy debate signals that Chinese officials view AI-driven displacement as a structural challenge requiring fundamental rethinking of employment policy, not just incremental adjustments.

In Washington, China is often assumed to be pursuing aggressive and unencumbered AI diffusion, unconstrained by the democratic accountability or regulatory debates that complicate growth in the United States. The reality is far more complicated. Beijing’s drive to diffuse AI will increasingly run up against its commitment to employment stability and fear of collective action. Whether this ultimately slows diffusion remains to be seen. But the structural tension between these two mandates is real and growing. It implies that China’s political system may be structurally less capable of disruptive AI diffusion than many assume.