Chinese Courts Rule That Firing Workers to Replace Them With AI Is Illegal

In a ruling that has no precise parallel anywhere else in the world, Chinese courts have established that companies cannot legally terminate employees simply because an AI system can now perform their jobs more cheaply.

The decision comes from the Hangzhou Intermediate People’s Court, which on April 28, 2026, published a compendium of “typical cases” involving AI firms and worker protections — a deliberate act of legal signaling designed to set expectations across China’s technology sector. The most closely watched case involves a quality-assurance engineer referred to in court documents as Zhou, whose case has become the defining example of how Chinese labor law applies to AI-driven restructuring.

The Case That Built the Precedent

Zhou’s employer — a tech company whose name has been withheld — reduced his monthly salary from 25,000 yuan (approximately $3,450) to 15,000 yuan after deploying a large language model to handle the bulk of his quality-assurance responsibilities. When Zhou refused to accept the pay cut, the company terminated his contract, citing “objective major changes” that made the original employment terms impossible to fulfil.

Under Chinese labor law, “objective major change” is one of the few legitimate grounds for terminating a contract without the employee’s consent. The company’s argument was that the integration of AI into its workflow constituted just such a change — a position that, if accepted, would have opened the door to AI-driven layoffs across the sector.

The Yuhang District Court rejected that argument, and the Hangzhou Intermediate Court upheld it on appeal. The ruling held that AI adoption and the cost savings it generates do not constitute a “major objective change” in the legal sense — the change, the court found, was a business decision by management, not an external condition that made the contract impossible to perform. The company was ordered to pay Zhou back wages and legal costs.

A parallel ruling from a court in Shenzhen, published in December 2025 and reinforced in the April 28 compendium, reached similar conclusions: dismissals motivated primarily by the desire to reduce headcount through AI substitution are unlawful absent demonstrated financial distress, genuine business closure, or documented individual performance failures.

Why This Matters Beyond China

The international significance of the ruling is not that Chinese companies will suddenly stop deploying AI — they will not, and the courts have been careful not to suggest that AI adoption itself is problematic. The significance is the legal distinction the courts have drawn: AI deployment is a business choice, and the costs of that choice cannot be externalized onto workers through unilateral contract termination.

No equivalent protection exists in the United States, where at-will employment law gives employers broad latitude to restructure workforces for any economically rational reason, including AI substitution. The European Union’s forthcoming AI Act contains provisions around “high-risk” AI systems in employment contexts, but enforcement mechanisms remain weak and the Act does not create a direct right to reinstatement for workers replaced by AI.

The contrast is stark. In Q1 2026, global tech layoffs tied explicitly to AI automation exceeded 78,000, according to data compiled by The Next Web. Companies including Snap, IBM, Salesforce, and various mid-market software firms have all cited AI efficiency gains as a factor in headcount reductions over the past six months, typically without legal challenge.

Chinese workers now have a legal lever that their counterparts elsewhere do not.

The Enforcement Landscape in China

The practical question is how aggressively these rulings will be enforced. China’s labor law system depends heavily on workers filing formal complaints or lawsuits — a high-friction process that many employees, especially in precarious positions, are reluctant to use.

Still, the publication of “typical cases” by the Hangzhou court is the Chinese legal system’s equivalent of issuing regulatory guidance. It signals to HR departments and in-house legal teams what the courts will and will not accept. Companies that attempt to replicate Zhou’s dismissal scenario are now on notice that they face litigation risk, even if enforcement remains uneven.

The rulings also arrive in the context of broader Chinese government concern about youth unemployment, which hovered near 14% in early 2026, and structural economic shifts that have put significant pressure on mid-skilled white-collar workers in the technology sector. The court decisions align with official policy direction, which makes them more likely to be followed consistently in lower courts.

What It Means for Global AI Policy

The ruling has drawn attention from labor law scholars in Europe and the United States who are tracking whether analogous protections should be built into their own frameworks.

“China has moved faster here than any Western jurisdiction,” said one labor policy researcher quoted by The Next Web. “They’ve essentially said: the efficiency gains from AI belong to the employer, but the losses cannot automatically be transferred to the worker. That is a coherent legal principle.”

Whether that principle travels to other legal systems depends heavily on political dynamics that differ sharply by country. In the US, where corporate lobbying on workforce matters is extremely effective and labor law reform is perennially stalled, the near-term prospects for similar protections are limited. In Europe, where worker council requirements and social partnership models give labor more legislative leverage, the Hangzhou ruling may find a more receptive audience.

For the global AI industry, the meta-story is a familiar one: technology that is deployed faster than regulation can keep pace eventually generates legal backlash. The Hangzhou court’s ruling is an early, and unusually specific, instance of that backlash taking formal legal shape.

The question now is not whether other jurisdictions will follow — it is when.