Skip to content
Welcome, thoughtbot's Giant Robots listeners!Start here
AI: Jobs, Power & Money
10APR

Hangzhou court bans AI-driven worker sackings

4 min read
16:54UTC

Hangzhou Intermediate People's Court released two judgments on 28 and 30 April 2026 ruling AI-driven dismissal illegal without prior retraining and reasonable reassignment, the first binding worker protection of its kind anywhere in the world.

EconomicDeveloping
Key takeaway

China is the first jurisdiction with a binding legal route against AI-driven dismissal, and its courts moved in days.

Hangzhou Intermediate People's Court released two judgments on 28 and 30 April 2026 upholding rulings that AI-driven dismissal is illegal without prior retraining and reasonable reassignment, under China's Labor Contract Law⁠1. The Yuhang District Court found in favour of Zhou, a quality assurance supervisor at an unnamed Chinese tech company who earned 25,000 yuan a month verifying AI-generated content. When the company's AI took over his duties, it offered him a new role at 15,000 yuan, a 40 per cent pay cut, and terminated him for refusing, paying 311,695 yuan in severance and citing "organisational restructuring"⁠2. The intermediate court upheld the ruling, then released both judgments together on the eve of International Workers' Day.

The legal test that matters is whether AI cost savings count as a "major change in objective circumstances", the clause under the 2008 Labor Contract Law that allows employers to break employment contracts. Yuhang ruled they do not. That makes deliberate automation a strategic choice the employer must absorb the consequences of, not an external shock the worker must wear. China's Ministry of Human Resources and Social Security recognised 42 new AI occupations in April 2026: the planning route. The Hangzhou and Beijing judgments now sit alongside it as a parallel judicial route, and the deliberate Workers' Day release confirms the political signal that Chinese courts are not independent of state policy in the way Western readers understand judicial independence.

For readers in San Francisco, London or Berlin, the immediate effect is zero: no Western jurisdiction has equivalent protection. Microsoft, Apple or Tesla restructuring its China operations now bears a per-head termination cost it does not bear at home, narrowing the cost gap between offshoring and reshoring labour-intensive functions. Multinationals operating inside China face a precedent that can be cited in any Chinese labour court tomorrow.

The doctrinal move, procedural protection before substantive statute, has a historical echo in Weimar Germany's 1920 Works Council Act, which codified worker consultation rights inside firms a decade before any general dismissal law. Beijing has produced any binding answer at all, which is more than Brussels or Washington can say this week.

Deep Analysis

In plain English

In most countries, if your employer decides a machine can do your job, they can let you go. China's Hangzhou Intermediate Court ruled in April 2026 that employers must first attempt retraining before any AI-driven dismissal becomes lawful. Two courts in Hangzhou ruled in April 2026 that a company planning to use AI must first try to retrain affected workers and find them other roles. Only if both those things fail can it dismiss someone. The ruling used China's 2008 employment law, which requires a genuine, unforeseeable business change before a firing is legal. The courts said choosing to automate is a deliberate business decision, not an unexpected event, so the employer carries the cost. The practical result: any employer in China with AI restructuring plans now needs a documented retraining programme or faces legal exposure.

Deep Analysis
Root Causes

China's Labour Contract Law Article 40(3) permits dismissal when 'objective circumstances change materially, making it impossible to fulfil the original contract.' Courts had for 15 years interpreted 'objective' broadly, covering market shifts, client loss, and product discontinuation. Employers began filing AI restructuring under this provision after 2023.

The Yuhang District Court's innovation was to apply the 2020 Supreme People's Court guidance on 'subjective versus objective causation' in contract frustration, ruling that a planned, budgeted AI system rollout is inherently subjective (the employer's choice), not objective (external). The Hangzhou Intermediate Court's endorsement elevated this to a two-court chain sufficient to be cited in similar future cases.

A secondary cause: China's Ministry of Human Resources simultaneously recognised 42 new AI-related occupations in April 2026, creating an official occupation taxonomy that implicitly assumes retraining is feasible. Courts noting this policy context have a stronger basis for imposing retraining obligations.

What could happen next?
  • Consequence

    Multinational employers with China operations will add mandatory retraining documentation to AI deployment projects, increasing timelines by 12-18 months for roles covered by Chinese Labour Contract Law.

    Short term · 0.75
  • Precedent

    South Korean and Japanese courts, which apply structurally similar 'objective circumstances' tests in employment law, are likely to cite the Hangzhou standard in domestic AI-dismissal cases by end 2026.

    Medium term · 0.6
  • Risk

    Employers who execute nominal retraining programmes without genuine reassignment intent will face follow-on litigation as workers and unions test whether courts distinguish genuine from pro-forma compliance.

    Medium term · 0.7
First Reported In

Update #8 · Beijing court bans AI sackings as Big Tech burns cash

Xinhua· 2 May 2026
Read original →
Different Perspectives
Salesforce, Synopsys and TD Bank Group
Salesforce, Synopsys and TD Bank Group
Salesforce, Synopsys and TD Bank Group each filed quarterly disclosures in late August booking restructuring charges, or none at all, without naming AI as a cause. Their silence matters because Challenger's tracker shows AI as a stated reason fell to fourth place in August even as the year-to-date AI-cut total still leads at 116,175.
Singapore, South Korea, Taiwan and Indonesia
Singapore, South Korea, Taiwan and Indonesia
Singapore launched its Skills and Workforce Development Agency on 16 September, giving citizens six months of free premium AI tools, while South Korea ring-fenced its AI tax windfall in a new Future Response Fund. Taiwan kept funding its AI build past NT$190bn and Indonesia rewired vocational training around AI literacy, betting state-built skills beat a market-led adjustment.
ver.di, CGT Fonction Publique and CCOO
ver.di, CGT Fonction Publique and CCOO
Germany's ver.di banked a 3.3% pay rise on 1 September and opened talks on a Tarifvertrag Transformation covering dismissal bans and reskilling, while France's CGT rejected Paris's AI negotiating timetable the same week. Spain's CCOO went further on 21 September, proposing to tax companies by the jobs they generate rather than wait for the next bargaining round.
BIS General Manager and Federal Reserve governors
BIS General Manager and Federal Reserve governors
The BIS's General Manager said on 10 September that AI displacement remains limited, even as the BIS's own survey found nearly 80% of firms plan to automate roles. Two Federal Reserve governors made the same point in July, arguing the labour-market data does not yet show a mass-firing event.
Bank of Canada, ONS and ECB
Bank of Canada, ONS and ECB
The Bank of Canada found the job-finding gap between AI-exposed and unexposed occupations widened from 2.2 to 13.9 percentage points since 2015-19, while separations barely moved. That framing, a hiring freeze rather than a firing wave, is echoed by the ECB's finding that euro-area AI use hit 52% of workers in 2026, concentrated among the university-educated.
Office for National Statistics
Office for National Statistics
Deferred its Transformed Labour Force Survey beyond November 2027 and disclosed a May 2026 telephone-collection failure. The ONS carries no AI-attribution layer at all, so Britain sits outside this month's cohort of measuring states by its own admission.