A Chinese tech company wanted to replace one of its employees with artificial intelligence. The AI was cheaper. The math was simple. The law, it turned out, was not.
The company cut the worker's monthly salary from 25,000 yuan (roughly $3,450 USD) to 15,000 yuan - a 40% reduction - citing the fact that AI could now do his job at a lower cost. When the worker, identified only as Zhou in court documents, refused to accept the pay cut, the company fired him. Zhou challenged the dismissal. He won at arbitration. The company sued anyway. It lost again at a district court. Then it appealed to the Hangzhou Intermediate People's Court. It lost a third time.
The ruling, published on April 28, 2026, was timed deliberately - one day before International Workers' Day. It arrived alongside six other labor dispute rulings from Hangzhou courts, all involving AI companies and their employees. Together, they form something unusual in the history of modern automation: a legal framework, issued in real time, telling a booming industry exactly how far it can go.
The background
Hangzhou is not just any Chinese city. It is the home of Alibaba, one of the world's largest technology companies, and has spent years positioning itself as China's capital of artificial intelligence. The city's courts have a specific mandate - to serve what authorities describe as the goal of making Hangzhou the "number one city for AI innovation and development."
China's core AI industry - meaning the companies and services that directly develop or deploy artificial intelligence - exceeded 1.2 trillion yuan in 2025, with more than 6,200 related enterprises operating across the country. That is equivalent to roughly $165 billion USD, or more than the entire annual economic output of countries like Hungary or Ukraine. The industry is growing fast, and the workers inside it are multiplying just as quickly: China graduated over 120,000 AI-related majors in 2024, a 35% increase year-over-year.
But growth in an industry does not automatically mean security for the people working in it. The opposite can be true. As AI tools become more capable, companies face a specific temptation: use the technology not just to expand what they do, but to shrink what they spend on labor. A worker who once took hours to do a task may find that an AI system does the same task in seconds. The company saves money. The worker loses income - or the job entirely.
In China, this tension runs into a particular legal obstacle: the Labor Contract Law, a piece of legislation that governs when and how employers can terminate workers. Under the law, a company can legally end a contract if there has been a "major change in objective circumstances" that makes the contract impossible to perform. The phrase sounds technical, but its meaning matters enormously. Courts have historically interpreted it to cover things like natural disasters, government-ordered closures, or company mergers - genuine disruptions from outside. The question now was whether a company's own deliberate decision to automate counted as such a disruption.
What is actually happening
Zhou worked at an unnamed tech company in Hangzhou doing question quality inspection - reviewing interactions between AI systems and users to verify the accuracy of the AI's answers. His job, in other words, was to check the AI's work. Then the company decided it no longer needed a human to do that.
At the heart of the case was whether AI-driven job replacement constitutes a "major change in the objective circumstances" that can lead to termination under China's Labor Contract Law. The intermediate court found that the grounds the company cited did not constitute such a "major change," which typically refers to significant events like company relocations or mergers. It also ruled that the company had failed to demonstrate that the contract had become impossible to perform. The alternative position offered to Zhou came with a substantial pay cut, which the court ruled was not a reasonable reassignment proposal.
The logic matters. The company argued, essentially, that market forces had changed around it - that AI had become so capable and cost-efficient that continuing to employ Zhou made no economic sense. The court rejected this framing. A company choosing to automate is making a business decision. Business decisions are not natural disasters. They do not, by themselves, dissolve a legal obligation to an employee.
A Zhejiang lawyer not connected to the case told state-run news agency Xinhua that AI adoption does not automatically justify a company terminating a labor contract to cut costs.
This case did not arrive in isolation. A Beijing case came first. In late 2024, a tech company eliminated the department of a map data collector named Liu after switching to AI-powered automated data collection. The company terminated Liu's contract, citing "major changes in objective circumstances." Liu challenged the firing. In December 2025, the Beijing Municipal Bureau of Human Resources and Social Security published the case as a model ruling for the year, finding the AI pivot was a deliberate, predictable strategy - not an unforeseeable disruption - and therefore the dismissal unlawfully shifted the risks of technological change onto the employee.
The Hangzhou ruling on April 30 followed the same line of reasoning. The two cases, from two different cities, now point in the same direction.
The money trail
The economics of this moment are not subtle. Since the start of 2026, tech companies have announced 78,557 job cuts worldwide, of which about 76.7% were by U.S.-based firms. Baidu ended 2025 with nearly 7% fewer employees, while electric-vehicle maker BYD cut roughly 10% of its workforce. AI is not the only explanation - Chinese experts point to broader pressures such as a slowing economy, weak consumer demand, and a prolonged property crisis - but it provides a convenient narrative cover.
There is a specific financial logic at work. When a company replaces a worker with an AI tool, it shifts a recurring labor cost (salary, benefits, legal obligations) into a different kind of expense (software licensing, compute costs, training data). The recurring cost is often higher on paper. The AI option scales more cheaply. For companies under margin pressure, the math is attractive.
But the Hangzhou ruling reframes that calculation. If an AI replacement does not constitute valid legal grounds for dismissal, then a company that fires a worker citing automation may owe that worker double severance pay - the penalty under Chinese law for wrongful termination. The exact figures in Zhou's case have not been publicly disclosed, but the precedent means that every similar dismissal now carries legal and financial risk.
Legal scholars have emphasized a key principle: the costs of technological transformation should not be borne solely by workers. Companies should not use AI adoption as a pretext for layoffs or as a means to sidestep their obligations. Pan Helin, an economist and member of an expert committee under China's Ministry of Industry and Information Technology, argued that while AI-driven job displacement may be inevitable, companies must ensure fair treatment during transitions, including reasonable reassignment arrangements and adequate compensation.
The six other cases published alongside Zhou's tell a parallel story. A robotics company employee named Zhao was convicted of stealing trade secrets, sentenced to prison - and still had to pay back his non-compete compensation and a separate fine to his former employer. A different company tried to recover 500,000 yuan in source code costs from a full-stack engineer it had fired during probation; the court ruled that software licenses purchased before an employee starts are a standard business cost, not the worker's liability. A third case established that companies cannot refuse to pay overtime simply because the overtime was not pre-approved through an internal form - if the work happened, the pay is owed.
Taken together, these rulings draw a fairly clear line. Companies in China's AI sector cannot offload the costs of their own strategic decisions - whether automation, failed hires, or software purchases - onto individual workers.
What people are doing about it
Workers across China have been watching closely. The case is among several labor disputes arising from AI job replacements across Chinese cities, and trade media has covered the Zhou ruling extensively. Chinese social media platforms saw the ruling trend in the days before May 1 - whether by design or coincidence, the timing amplified the signal.
China's 2026 government work report called for improving measures to promote employment and entrepreneurship in response to the development of AI, marking the inclusion of AI's impact on jobs within a national policy framework for the first time. That framing matters. Beijing is simultaneously pushing companies to adopt AI more aggressively and signaling to workers that the law still protects them - a tension that the courts are now being asked to resolve, one case at a time.
A group of researchers from Peking University and Renmin University of China estimated in 2020 that AI could displace up to 278 million Chinese workers by 2049, equivalent to roughly a third of those currently employed. Another study from 2023 estimated that 54% of jobs are at high risk of substitution in the coming decades, with routine processing work first to go. Against that backdrop, individual court rulings are limited instruments. But they establish norms, and norms shape behavior - particularly for companies concerned about reputational risk with a government that has made employment stability a stated priority.
The Chinese government has pledged to create more than 12 million urban jobs in 2026 and keep the urban unemployment rate below 5.5% through 2030. Companies that visibly shed staff in the name of AI cost-cutting are doing so with an audience that includes regulators.
At the same time, companies are not standing still. BYD, which laid off part of its frontline production and sales staff, accelerated hiring for research and development roles by nearly 5% in 2025. The pattern is familiar from other economies: automate the old jobs, hire for the new ones. The question is whether the people holding the old jobs can become the people needed for the new ones - and how much time, money, and legal protection they have to make that transition.
The bottom line
What Hangzhou's courts established this week is a simple but consequential idea: a company's voluntary decision to adopt a cheaper technology is not a natural disaster, and it does not cancel its legal obligations to the humans it hired. The ruling does not stop automation. It does not prevent companies from restructuring. It says, very specifically, that using AI as a pretext to fire workers without proper compensation is illegal - and that cutting pay by 40% as a condition of keeping a job is not a reasonable offer. In an industry expanding at the pace China's AI sector is growing, courts in the city at the center of that expansion have just told every company in it: the cost of progress is not something workers are required to absorb alone.
Timeline
- July 2009 - Liu begins work as a map data collector at a Beijing tech company
- Early 2024 - The Beijing company switches to AI-powered data collection, eliminates Liu's department, and terminates his contract
- Late 2024 - Liu wins arbitration, challenging the dismissal as unlawful
- December 26, 2025 - Beijing Municipal Bureau of Human Resources and Social Security publishes Liu's case as a model ruling, finding the AI pivot was a deliberate business strategy, not an unforeseeable disruption
- 2025 - Zhou, a question quality inspector at a Hangzhou tech company, is offered a reassignment with a salary reduction from 25,000 yuan to 15,000 yuan; he refuses and is dismissed
- 2025 - Zhou wins at arbitration; the company files a lawsuit in Hangzhou's Yuhang District Court and loses
- August 2025 - The company appeals to the Hangzhou Intermediate People's Court
- April 28, 2026 - Hangzhou Intermediate People's Court publishes seven typical labor dispute cases involving AI companies
- April 30, 2026 - Court upholds Zhou's wrongful dismissal ruling, finding the dismissal unlawful on two counts
- May 1, 2026 - International Workers' Day; ruling receives widespread coverage internationally
Summary
Who: An unnamed tech worker referred to as Zhou, employed by an unnamed Hangzhou AI company; the Hangzhou Intermediate People's Court; and China's broader AI industry workforce.
What: A Chinese court ruled that a company cannot legally fire an employee simply because AI can perform the same role at lower cost. A 40% forced pay cut was ruled an unreasonable reassignment offer; the dismissal was deemed unlawful.
When: The ruling was published April 28-30, 2026, alongside six other AI labor dispute cases from Hangzhou courts.
Where: Hangzhou, Zhejiang Province, China - the country's leading AI industry hub and home to Alibaba.
Why: As China's AI sector surpasses 1.2 trillion yuan in size and companies begin citing automation as grounds for layoffs and pay cuts, courts are establishing the legal limits of what employers can do when technology displaces human roles.