TL;DR
On September 7, 2026, China's Supreme People's Court (SPC) issued the Opinion on Properly Hearing Cases Involving Artificial Intelligence Disputes (Fa Fa [2026] No. 10) — 5 parts and 24 articles in total, the first judicial ruling guideline on AI issued by the country's highest judicial body, setting out liability rules for scenarios such as AI face-swapping and voice cloning, online doxxing, autonomous-driving accidents and AIGC copyright infringement.
Background
The Opinion was released by the SPC at a press conference on September 7, 2026, under document number Fa Fa [2026] No. 10. It is the first AI-related judicial ruling guideline issued by the highest judicial body in China. Since China has not yet enacted a dedicated AI law, the Opinion is based on existing statutes — the Civil Code, Cybersecurity Law, Data Security Law, Copyright Law, Anti-Unfair Competition Law, Consumer Rights Protection Law, Personal Information Protection Law and Civil Procedure Law — providing courts at all levels with guiding principles and rules for hearing AI-related disputes. It is problem-oriented, targeting widely discussed issues such as "AI face-swapping and voice cloning," "AI hallucination" infringement, "online doxxing," "big-data price discrimination," autonomous driving and model training; issues on which consensus is not yet possible are deliberately left open.
1. Overall framework: 5 parts, 24 articles
| Part | Articles | Main content |
|---|---|---|
| Part 1 | 1–2 | Guiding ideology and three basic principles (people-centered, support innovation, secure the safety baseline) |
| Part 2 | 3–11 | Tort cases: attribution principle, personality rights, personal-information rights, consumer rights, generative-AI provider liability, AI product liability, compensation for traffic accidents involving autonomous and assisted driving |
| Part 3 | 12–16 | IP: infringement liability, open-source software liability, patent grant and validity, technology contracts, data-use norms |
| Part 4 | 17–20 | Procedure: fact-finding and evidence review, curbing improper evidence collection; principled rules on punishing AI-enabled crimes |
| Part 5 | 21–24 | Working mechanisms: dispute prevention and resolution, research, case guidance, trial supervision, inter-agency coordination and international judicial cooperation |
2. Core tort-liability rules
Attribution principle (Article 3): where the law does not expressly provide for no-fault or presumed-fault liability, fault liability under Article 1165(1) of the Civil Code applies; fault is assessed by weighing the AI application scenario, degree of autonomy, technical and information transparency, potential risk and scope of impact, plus the measures taken and technical feasibility for preventing harm.
Personality rights (Article 4): using AI to process a natural person's name or likeness and generating an identifiable virtual digital image of that person without consent constitutes infringement of the right to name and right to likeness; using a person's voice as training data without consent to produce identifiable synthetic speech infringes voice rights; manipulating a virtual image or synthetic voice to misbehave or publish false statements that lower another's social evaluation constitutes infringement of the right to reputation; unauthorized creation or use of a deceased person's virtual image that harms the deceased's name, likeness or reputation entitles close relatives to claim liability under Article 994 of the Civil Code.
Privacy (Article 5): using AI to track and analyze a person's publicly available information (phone numbers, online accounts, etc.) with the intent of prying into privacy, obtaining private information and leaking or publicizing it constitutes infringement of privacy. The Opinion also addresses whether training models on publicly available personal information constitutes infringement, the definition and defect determination of AI products, and compensation for traffic accidents involving autonomous and assisted-driving vehicles.
3. Intellectual-property rules
Articles 12–16 specifically govern AI-related IP disputes: where AI-generated content is accused of infringing copyright, the relevant party cannot evade liability merely because the content was generated by AI; liability must be proportionate to control capability and duty of care. Courts weigh the type of AI service, industry characteristics, training-data sources, degree of each party's participation, necessary measures taken and profits gained to determine the liability of developers, providers and users. Using AI for infringement and counterfeiting, false advertising, or fake traffic and transaction manipulation that harms others' rights or constitutes unfair competition incurs liability under law. Articles 12–16 also cover open-source software liability, patentability of AI-assisted inventions, technology-contract performance and data-use norms.
4. Procedure and governance mechanisms
Procedure (Articles 17–20): given the technical nature of AI cases, the Opinion develops fact-finding and evidence-review rules, regulates improper evidence collection and submission of false materials that obstruct judicial order, and sets principled rules on punishing AI-enabled crimes under a policy of leniency combined with severity — deliberate use of AI to endanger national security, infringe citizens' rights or disrupt social order is strictly punished, while innovative conduct in AI R&D and application is handled cautiously under the principles of legality and proportionality.
Working mechanisms (Articles 21–24): provisions on preventing and resolving AI-related disputes, research and evaluation, case guidance and trial-supervision guidance, inter-agency coordination and international judicial exchange, reflecting that AI governance is a systematic project requiring multi-party coordination.
Common misconceptions and details
- The Opinion is not legislation but judicial ruling guidance. It creates no new legal obligations; it unifies adjudication standards within the existing legal framework, and issues lacking consensus are deliberately left open.
- "AI-generated content has no copyright" is not the Opinion's conclusion. It sets liability-allocation rules — when AI-generated content infringes copyright, the liability of developers, providers and users is apportioned in line with control capability and duty of care.
- Autonomous-driving accidents are not automatically blamed on automakers. Courts determine the liable party and compensation based on existing law and the specific scenario.
Summary
The SPC Opinion on Properly Hearing Cases Involving Artificial Intelligence Disputes builds China's first systematic judicial framework for AI civil disputes in 5 parts and 24 articles, covering personality rights, personal information, consumer rights, intellectual property, competition order and procedure. It establishes the principles of people-centeredness, innovation support and safety baseline, while leaving room for future legislation and evolving practice.
Sources and references
- SPC official website: People-centered, AI for good — focus on the Opinion: link
- SPC (full text): SPC issues the Opinion on Properly Hearing AI Disputes: link
- Chinacourt: Balancing personal-information rights and AI industry development: link
- JCRB: SPC clarifies liability for AI-based infringement of personality rights: link
- Guangming: SPC clarifies adjudication rules for "AI face-swap and voice-clone" cases: link
- Sina Finance: In-depth reading of the Opinion: link