AI liability and pending legislation in China
China has no dedicated AI law yet and no AI-specific civil liability instrument. Harm from AI is handled by general civil law plus administrative rules — and comprehensive AI legislation is now an explicit item on the State Council's 2026 legislative plan.
In one line
Liability today is administrative-first: regulators act through filing, assessment, takedown and fines; private claims fall back on the Civil Code. A comprehensive AI law is now formally on the legislative agenda.
Comprehensive AI legislation — status
This is a status note, not a law card. No AI law exists yet, so
publishedandeffectiveare empty by design: there is no instrument to date. The routes below are the ones currently in force.
- The State Council 2026 Legislative Work Plan (issued 11 May 2026) directs: accelerate comprehensive legislation for the healthy development of artificial intelligence, and improve legislation on AI's shared elements — data, computing power, algorithms, intellectual property, network security, supply chain security — and on key application scenarios.
- The framing has moved from "study" to comprehensive legislation, but no bill text has been published. Treat any date as speculative ⚠.
Where liability currently sits
| Route | Instrument | Notes |
|---|---|---|
| Civil — tort | Civil Code, tort liability chapter (effective 2021-01-01) | Fault-based liability, with special rules for specified situations. |
| Civil — product | Civil Code, product liability provisions | Applies to defective products; whether an AI-enabled service counts as a "product" is unsettled ⚠. |
| Statutory AI clause | Cybersecurity Law, new Article 20 (amendment adopted 2025-10-28, in force 2026-01-01) | Supports basic AI research, training-data and computing infrastructure, AI ethics and risk monitoring, and the use of AI to raise network-security protection. A promotional clause — it creates no liability of its own. |
| Administrative | Interim Measures for Generative AI Services (effective 2023-08-15) | Filing, security assessment, content compliance; penalties handled under the Cybersecurity Law, Data Security Law, PIPL and similar. The amended Cybersecurity Law (in force 2026-01-01) reaches up to RMB 10 million, with up to RMB 1 million on responsible individuals, on two separate triggers — Article 61 for network-security failures and Article 69 for prohibited information — each transcribed in the Cybersecurity Law card. |
| Administrative | Algorithmic Recommendation Provisions (Order No. 9) | Filing, security assessment, warnings, suspension of information updates, RMB 10,000–100,000 fines. |
| Ethics | Measures for Ethical Review of Science and Technology (trial) (effective 2023-12-01) | Ethics review as a pre-condition for certain research and applications. |
What is not there yet
- No statutory definition of AI harm or of the liable party across the chain (developer, provider, deployer, user).
- No dedicated burden-of-proof rule for AI-related damage — the gap the withdrawn EU directive was meant to fill.
- No published draft, so the eventual scope (registration? liability allocation? model-level duties?) is unknown.
Unverified / TODO
- Confirm the Generative AI Interim Measures' penalty article and cross-references — Article 21: punishment follows the Cybersecurity Law, Data Security Law, Personal Information Protection Law and Science and Technology Progress Law; where those are silent, warning, circulated criticism and an order to correct within a time limit, and on refusal or serious cases an order to suspend the relevant service. No fine amount of its own. See
generative-ai-services-measures.md - Check whether the Supreme People's Court has issued AI-related judicial guidance ⚠
- Track the comprehensive AI law from plan to draft and note the first public consultation [gap:no_instrument]
- Confirm the Civil Code article numbers for product liability
Official sources
- State Council 2026 Legislative Work Plan (MOJ): https://www.moj.gov.cn/pub/sfbgw/zwgkztzl/xxxcgcxjpfzsx/fzsxyw/202605/t20260511_534688.html
- Xinhua report on the plan: https://www.news.cn/20260511/64fb4f5178b042e6aa16bf55932f038e/c.html
- Cybersecurity Law, re-promulgated text (CAC): https://www.cac.gov.cn/2025-12/29/c_1768735112911946.htm
- Personal Information Protection Law (NPC): http://www.npc.gov.cn/npc/c2/c30834/202108/t20210820_313088.html