Chinese Court Precedent Prohibits Layoffs on Grounds of AI Substitution, Sparking Debate Between Employment Protection and Innovation Limits

A second-instance ruling by the Hangzhou Intermediate People's Court concluded that an employer's actions—reassigning an employee, reducing salary, and subsequently dismissing them on the grounds that their work could be replaced by AI—violated the law, and ordered the employer to pay over 260,000 yuan in compensation. Earlier, the Guangzhou Intermediate People's Court issued a similar ruling in a case involving a graphic designer position replaced by AI. Judicial practice clarifies that technological iteration does not constitute a lawful basis for employment adjustments.

Cause and Operational Mechanism

The plaintiff, Mr. Zhou, served as a supervisor. He was reassigned with a salary cut and eventually dismissed due to the alleged replaceability of his work by AI. The court, citing relevant provisions of labor law, ruled that the employer failed to demonstrate that it could not reallocate the employee through alternative positions. The Beijing Municipal Human Resources and Social Security Bureau subsequently issued guidance requiring employers to assess the possibility of reassignment, provide training, and extend notice periods by at least 90 days before conducting automation-related layoffs. Employers must establish evidentiary records proving that positions are technologically or economically unsustainable to maintain.

This mechanism stems from Article 40 of the Labor Law, which strictly regulates economic layoffs. The court emphasized that productivity improvements do not constitute sufficient grounds for terminating employment contracts, and introduced considerations of occupational dignity, thereby increasing grounds for compensation. Non-compliant employers face potential fines of up to 50% of annual turnover.

Impacts on Stakeholders

For workers, these precedents offer protection to 200 million employees potentially threatened by intelligent automation, reducing post-incident legal costs, while requiring continuous skill upgrades to adapt to human-machine collaboration models. For employers, the previous path of directly cutting labor costs through AI is now constrained, forcing investment in training and reassignment, which increases operational burdens in the short term but may foster more sustainable human-machine collaboration over time.

Developers and AI tool providers must adjust product positioning from replacement tools to augmenting human capabilities. Enterprise users in industries such as finance, logistics, customer service, and administrative management will face stricter pre-deployment assessments, needing to demonstrate that alternative options have been exhausted before implementing AI.

International Comparisons and Policy Differences

China's approach contrasts with models in the U.S. and U.K., which encourage creative destruction. China prioritizes social stability over short-term productivity gains, rooted in balanced development philosophy. The EU is studying requirements for similar reassignment proof within its AI Act, Japan announced a similar draft law in January 2026, and South Korea experiments with tax incentives in technology transition zones.

Chinese tech companies have begun establishing career transition programs. Alibaba trains employees to become algorithm supervisors, while Tencent's internal academy has reduced layoffs. Microsoft's experience in the Chinese market has also shifted toward AI-assisted reassignment rather than mass layoffs.

Forward-Looking Assessment

Based on existing precedents and guidance, the most likely development is that more localities will establish monitoring and early warning systems for AI-related employment impacts, requiring process monitoring before large-scale AI deployment by enterprises. Historical experience shows that sharing gains from technological change depends on institutional design. If China's existing labor law system incorporates new circumstances through judicial interpretations, it can impose constraints at the front end of the chain, avoiding inefficiencies in end-of-line remedies. Ordinary workers must simultaneously focus on skill improvement and protection of their rights regarding personal data and intellectual property collection.