Chinese court rules AI-only layoffs illegal

The whole ruling turns on a narrow legal question: does deploying AI count as an 'objective major change in circumstances' under Article 40 of China's Labor Contract Law? Article 40 was written to cover events imposed on a business from the outside — the kinds of shocks employers cannot reasonably plan for, like a forced relocation, a regulator-mandated shutdown, or a corporate merger. If a court accepts that an event qualifies, the employer can lawfully terminate workers whose roles can no longer be performed, subject only to statutory severance.

Chinese employers tried to fold AI replacement into that bucket, arguing that automation transformed the cost structure of certain roles so completely that retaining the worker was no longer feasible. The Hangzhou and Beijing courts rejected that framing on the same logic: the company chose to deploy AI. As the appellate reasoning in the Zhou case put it, 'A company's decision to adopt AI is a strategic business choice, not an unforeseeable change in objective circumstances.' The Beijing court was even more pointed about who absorbs the risk, calling these 'risks that the technology company should have been able to foresee in its normal operations.' The effect is to shut a procedural door: employers can still automate roles, but they cannot use Article 40 as the off-ramp for the worker.

The factual texture of the Zhou case matters more than the headline. Zhou had been a quality-assurance supervisor since November 2022, earning 25,000 yuan per month (about $3,655) to verify that the company's large language model output met content rules and filtered illegal material — a job that exists precisely because of AI, and that AI is now starting to do for itself. His employer's response was not a clean firing; it was a reassignment offer at 15,000 yuan a month, a roughly 40% pay cut. When Zhou refused, the company terminated him, which positioned the dispute as a worker rejecting a 'reasonable' alternative rather than as a layoff.

The courts saw through that structure and treated the reassignment-or-fire choice as a coerced termination dressed up in procedural clothing. Zhou was awarded 311,695 yuan in compensation. The signal to other employers is unusually concrete: an AI-driven internal demotion that materially cuts pay is treated as constructive dismissal, and the price tag for getting it wrong is more than a year's salary. Combined with the Beijing Liu precedent — a manual map-data role eliminated when the company switched to AI collection — the two cases now cover both ends of the playbook: outright firings and disguised demotions both fail the same legal test.

The numbers around the ruling tell a sharper story than the legal text. Reporting in TheNextWeb cites roughly 170,000 AI-cited layoffs in the United States that have faced no successful legal challenge, a function of at-will employment in 49 states (Montana being the lone holdout). Tom's Hardware tallies 78,557 tech layoffs from January through April 2026, with 76% in the US and 37,638 — 47.9% of the total — explicitly attributed to AI and workflow automation. In that environment, an American worker terminated in favor of a model has effectively no individual legal remedy; in China, after Hangzhou, that same worker has both arbitral and judicial routes to compensation.

The European picture sits in between. The EU AI Act, which becomes effective in August 2026, classifies AI used in employment decisions as high-risk and imposes documentation, transparency, and human-oversight obligations, but it does not prohibit AI-driven layoffs themselves. So the world's three major tech jurisdictions now occupy three distinct positions: the US treats AI replacement as an unconstrained business decision, the EU treats it as a regulated decision, and China treats the firing — though not the AI deployment itself — as presumptively unlawful. For multinationals, that means AI-driven workforce planning will increasingly need to be jurisdiction-specific rather than global.

The headline framing — 'China makes AI layoffs illegal' — overshoots what the courts actually did. AI replacement itself remains lawful; companies retain the right to deploy AI, restructure roles, and reduce headcount. What they lose is the right to invoke AI as a unilateral termination ground without negotiating, retraining, reassigning fairly, or paying full compensation. In Wang Xuyang's words, 'AI replacement does not automatically justify terminating a labor contract.' That is a procedural rather than substantive constraint, and a sufficiently capitalized employer can still automate aggressively — it just has to write the severance check.

There is also a sharper political read worth surfacing. Critics have pointed out that judicial independence in China is largely a fiction, and that two coordinated 'typical case' releases — Beijing in December 2025, Hangzhou in April 2026, the latter timed days before International Workers' Day — look more like regime-stability messaging than spontaneous jurisprudence. A minority commentator view argues the constraint may hand a competitive advantage to less-regulated jurisdictions, since Chinese firms now face higher headcount-reduction costs than US peers chasing the same automation gains. Whether the net effect is to slow Chinese AI deployment, accelerate it (because firms hire fewer humans to begin with), or simply redistribute its costs onto employers rather than workers is the open second-order question the Zhou and Liu rulings have not yet answered.