Sixty-one thousand. That is how many workers worldwide have lost their jobs to AI-related cuts since January 1, 2026, according to tracking data compiled by Blockonomi from layoffs.fyi and corporate filings. Nearly half of all tech layoffs this year, 47.9%, now cite artificial intelligence as a contributing factor. Meta alone is cutting 8,000 roles, roughly 10% of its workforce, with the savings redirected toward AI infrastructure spending that the company projects will reach $115 to $135 billion.
Into this accelerating vacuum of worker protections, a Chinese court just dropped a very large rock.
The Hangzhou Ruling
On April 30, the Hangzhou Intermediate People's Court upheld a lower court's decision that a tech company in Zhejiang province had illegally fired a quality assurance supervisor identified only as Zhou. His job: verifying the accuracy of answers generated by AI large language models. His situation is structurally ironic. Zhou earned 300,000 yuan per year (about $43,900) before the company decided its AI could do his work. It offered him a reassignment with a 40% salary cut, dropping him to roughly 180,000 yuan. He refused. They terminated his contract, citing the "disruptive impact of AI" on his role and "reduced staffing needs."
Zhou filed for arbitration. He won. Zhou's employer sued to overturn the arbitration. It lost at the district level. It appealed. It lost again.
Appellate reasoning was blunt: "The termination grounds cited by the company did not fall under negative circumstances such as business downsizing or operational difficulties, nor did they meet the legal condition that made it 'impossible to continue the employment contract.'" Adopting AI, the court concluded, is a voluntary business decision. It is not an earthquake, a pandemic, or a market collapse. You chose to deploy the technology; you cannot then treat the consequences of that choice as if they were an act of God that forces you to shed workers.
Not an Isolated Case
Hangzhou is not alone. In Beijing last year, a data mapping worker whose entire role was automated away also won his arbitration. Beijing's arbitration panel was strikingly precise: the company's decision to switch to AI was "a business choice rather than an uncontrollable event," and by terminating the employee, the firm was "shifting the cost of the technological transformation to the employee." Lawyer Wang Xuyang, commenting on the Hangzhou case to Xinhua, summarized the emerging legal consensus: "AI adoption doesn't automatically justify a company terminating a labor contract to cut costs."
The combined effect of these rulings is a legal framework that requires companies to retrain displaced workers, reassign them to comparable roles, or invest in building new skills rather than printing severance checks and calling it efficiency. Chinese labor contract law covers every formal employee in the country. There is no at-will exception.
The Global Comparison
What does the rest of the world offer workers facing the same predicament? Here is where the picture becomes stark.
| Jurisdiction | AI-Specific Worker Protection | Status | Enforcement Mechanism |
|---|---|---|---|
| China | Court rulings: AI replacement = illegal termination; must retrain/reassign | Active (case law) | Labor arbitration + civil courts |
| EU | AI Act classifies employment AI as "high-risk"; requires human oversight, transparency | Delayed to Dec 2027/Aug 2028 | Fines up to €35M or 7% global revenue |
| US Federal | None. WARN Act requires 60-day notice for mass layoffs (100+ workers) with no AI provision | No legislation pending | N/A |
| California (SB 951) | 90-day notice for AI layoffs affecting 25+ workers or 25% of workforce | Proposed | Back pay + civil fines |
| Minnesota (HF4369) | 90-day notice + transitional employment period + upskilling mandate | Proposed | Fines + loss of state tax credits |
| New York | Revised WARN form includes AI layoff disclosure checkbox | Active | Self-reporting (voluntary attribution) |
| Connecticut (SB 5) | Mandates AI-related workplace disclosures | Passed, effective Oct 2026 | To be determined |
The table's most revealing column is "Status." China has binding case law, right now, that a worker can invoke in court tomorrow morning. Brussels' protections exist on paper but keep getting pushed back, most recently via the Digital Omnibus package that delayed high-risk AI enforcement to late 2027 or mid-2028. American workers have a patchwork of proposed state bills, most of which haven't made it out of committee, sitting atop a federal void of precisely nothing.
The New York Experiment in Self-Reporting
New York offers the most instructive failure. In 2025, Governor Kathy Hochul made New York the first state in the country to require companies to disclose whether AI contributed to layoffs via a revised WARN Act filing form. It includes a checkbox. Simple enough.
And the result? According to Bloomberg Law's review, more than 160 companies filed AI-related layoff notices with the New York Department of Labor. Zero attributed the cuts to automation. Not one checked the box. Self-reporting produced a 0% attribution rate in a year when nearly half of all tech layoffs nationwide were AI-driven. If you build a speedometer that relies on drivers voluntarily reporting how fast they were going, you will conclude that no one speeds.
The Original Calculation: Coverage Gap
Here is a number nobody has published. Under current law, the federal WARN Act applies only to employers with 100 or more full-time workers and triggers only for layoffs affecting 50 or more employees at a single site, or 500 or more regardless of site, within a 30-day window. According to the Bureau of Labor Statistics, firms with fewer than 100 employees account for roughly 98.2% of all US business establishments and employ approximately 33.6% of the private-sector workforce, some 42 million workers, who fall entirely outside the WARN Act's reach even in its current non-AI-specific form. For the remaining 66.4% of workers at larger firms, the law requires only 60 days' notice with no retraining mandate, no reassignment obligation, and no prohibition on using AI replacement as grounds for termination. Contrast that with China's emerging framework, which covers every formal employee and requires companies to absorb the cost of their own automation decisions, could not be cleaner.
California's SB 951 and Minnesota's HF4369 would begin to close that gap, at least for workers in those states. Both mandate 90 days' notice instead of 60. Minnesota goes further by requiring a "transitional employment period" during which workers receive their salary while the employer provides upskilling opportunities. But both are proposals, neither has been signed into law, and even if they pass, they cover only their own residents. A software engineer laid off in Texas or Florida has the same federal protections in 2026 that she had in 1988, when the WARN Act was written: 60 days' notice if her employer is large enough, and nothing if it isn't.
The Strongest Case Against China's Approach
The most serious counterargument to China's labor protection model is that it could backfire in the hiring market itself. If companies cannot fire workers once AI makes their roles redundant, the rational response is to not hire humans for those roles in the first place. China's urban youth unemployment rate hit 21.3% in June 2023 before the government stopped publishing the metric entirely for several months. Protecting incumbent workers while making new hires riskier could create a two-tier labor market: secure insiders, locked-out newcomers. Some labor economists argue that Europe's strong employment protections have contributed to exactly this dynamic, with older workers protected by dismissal laws while youth unemployment in Spain and Greece exceeds 25%.
China's economy is also under sustained pressure from the rising costs of the Iran war's supply chain disruptions and a property sector that has not stabilized. Companies facing squeezed margins and mandated AI retraining costs may simply delay AI adoption itself, which could leave Chinese firms less competitive against American and European rivals that face no such friction. The question is whether that friction is a feature or a bug.
What This Analysis Does Not Show
Several limitations constrain this comparison. China's rulings come from city-level intermediate courts, not the Supreme People's Court. They are legally binding in their jurisdictions but do not yet carry the weight of national precedent in the way a US Supreme Court decision or an EU directive would. We do not have data on how many Chinese companies are quietly restructuring around these rulings by, for example, not replacing departing workers rather than formally laying them off. That 47.9% AI attribution figure for US tech layoffs relies on self-reported corporate disclosures, which New York's experience suggests dramatically undercount the real number. And the EU's enforcement timeline continues to shift, making it difficult to evaluate a regime that does not yet operationally exist.
The Bottom Line
A worker in Hangzhou who gets replaced by an AI chatbot can walk into court and win. A worker in Houston who gets replaced by the same technology can file for unemployment benefits. That is the gap, and it is widening every quarter as AI capabilities accelerate and legislative responses do not.
If you are a knowledge worker in the United States: check whether your state has pending AI layoff legislation and contact your state representative's office, because the federal government has shown no interest in acting. If you are an employer deploying AI: study the Hangzhou ruling carefully, because the legal principle that automation costs belong to the company that chose to automate, not the worker who was automated away, has a logic that travels across jurisdictions. And if you are a policymaker watching from Washington: 160 companies filed AI layoff notices in New York, and zero admitted it. Your voluntary frameworks are producing data that is indistinguishable from fiction. The world's second-largest economy just decided that is not good enough.
Sources
- NPR / Jennifer Pak (May 1, 2026). A tech worker in China is laid off and replaced by AI. Is it legal? whro.org (NPR)
- Blockonomi (May 1, 2026). China Makes AI-Driven Layoffs Illegal as Global Job Cuts Hit 61,000 in 2026. blockonomi.com
- Bloomberg Law (2026). AI-Related Layoffs Test New York's Ability to Track Job Losses. bloomberglaw.com
- HR Dive / Caroline Colvin (2026). New Minnesota bill would require 90-day notice for AI that could displace workers. hrdive.com
- Crowell & Moring (2026). Artificial Intelligence and Human Resources in the EU: a 2026 Legal Overview. crowell.com
- CGTN (Dec 27, 2025). New ruling for workers: Can you be fired if a job is replaced by AI? cgtn.com
- HR Grapevine (June 16, 2025). New York now requires companies to disclose AI-driven layoffs. hrgrapevine.com
- Bureau of Labor Statistics (2024). Business Employment Dynamics, establishment size data. bls.gov