Summary: Bloomberg reported on May 2, 2026 that the Hangzhou Intermediate People’s Court ruled an employer cannot dismiss a worker solely because AI can perform the role. The QA supervisor’s pay had been cut 40 percent during a proposed transfer, and both arbitration and the court found the dismissal unlawful. The article says it was the second such Chinese ruling in six months.
Bloomberg broke the news on May 2, 2026: the Hangzhou Intermediate People’s Court has handed down a landmark ruling — an employer may not dismiss a worker for the sole reason that AI can now perform that worker’s duties. The company in the case had concluded in 2024 that its own AI systems could take over a quality assurance supervisor’s role, moved him to a lower-ranked position while cutting his pay by 40 percent, and terminated him once he turned the transfer down. Both the labor arbitration panel and the court held the dismissal unlawful, and the company owes compensation.
What gives the ruling its weight is the classification: swapping a person out for AI is treated as a business call belonging to the employer, outside the “objective change of circumstances” that Chinese labor law accepts as grounds for unilateral termination. Within six months, this is the second time a Chinese court has reached the same conclusion, and the timing coincides with a global wave of tech layoffs carried out in AI’s name.
How the Dismissal Unfolded: 25,000 Yuan a Month, Then Out
The worker at the center of it is a QA supervisor surnamed Zhou, hired in November 2022. His daily work put him next to large language models — polishing their output and vetting sensitive content — at 25,000 yuan a month (about $3,640). Once 2024 arrived, the company judged that its AI systems could take on those tasks, moved him into a lower position, and dropped his monthly pay to 15,000 yuan. Zhou turned the new arrangement down and was dismissed on the spot. The labor arbitration committee found the firing unlawful; the Hangzhou Intermediate People’s Court upheld that finding and ordered the company to pay compensation. The decision went public this week, the second ruling of its kind to surface.
Where the Argument Turns: Article 40 of the Labour Contract Law
The same statutory hook holds up both rulings. Under Article 40 of the Labour Contract Law, an employer may terminate a contract when “objective circumstances” have changed so fundamentally that the contract can no longer be performed — wording historically reserved for external shocks such as force majeure or relocations the government ordered. The courts laid out their reading plainly: adopting AI is “a strategic business choice, not an unforeseeable change in objective circumstances.” A job made impossible by an external shock and a job made redundant by an internal decision are different things; only the first kind supplies grounds for termination. The cost of automation, put another way, may not simply be handed to the employee.
Six Months Before, Beijing Reached the Same Result
Beijing had already seen a nearly identical dispute in December 2025. The employee there, surnamed Liu, had collected map data by hand at a tech company from 2009 onward. Early in 2024, the company switched over entirely to AI-driven automated collection, dissolved its navigation products department, and dismissed Liu on the grounds of a major change in circumstances. The arbitration panel reasoned that introducing AI sat within the employer’s autonomous business decisions and its foreseeable operating risks. The Beijing Municipal Human Resources and Social Security Bureau went on to list the case among its ten most notable labor arbitration decisions of the year, and both the trial court and the appeals court let the ruling stand. Taken together, the two cases trace a settled judicial line in China: replacing someone with AI is not, on its own, a lawful basis for dismissal.
The View from the US and Europe
No Western country has a ruling to match. Employment in the United States is at-will in every state but Montana, so “replaced by AI” is not a forbidden reason to let someone go. A Senate bill that would have the Labor Department report quarterly on AI-related layoffs has gone nowhere; Illinois stops at requiring notice when AI factors into hiring, discipline, or discharge decisions; Colorado’s AI Act, which takes effect in mid-2026, demands risk-management policies rather than banning AI-based dismissals — and no US jurisdiction bars AI replacement alone as grounds for termination. The EU AI Act classifies AI used in recruiting, screening, and performance evaluation as high-risk, its obligations fully in force from August 2026 — yet the law governs how AI takes part in decisions, not whether companies may eliminate roles because of it. Union-championed protections, a European AI Social Compact among them, remain proposals on paper.
The Takeaway for Companies and Product Teams
Context gives this legal line its bite. Tech layoffs worldwide climbed past 78,000 in early 2026, roughly half attributed to AI, and Meta, Oracle, and Block all appear on the list; China’s urban youth unemployment, meanwhile, sits at 15.3 percent. The path Beijing has picked makes no move to restrict AI itself and instead pushes transition costs onto employers: to automate a role, a company must retrain the worker, reassign them at comparable pay, or keep them on. Klarna’s cycle of cutting customer-service staff and then rehiring them stands as a caution for technical teams — what AI tends to absorb is a bundle of tasks, not a whole position. For product and engineering teams operating in China, every AI deployment plan now carries a hard cost item with a court behind it: workforce transition, an obligation in law rather than a soft suggestion.