Case study
China’s Supreme People’s Court directs how existing law should handle AI disputes
What happened
The Supreme People’s Court issued Fa Fa [2026] No. 10 on September 7, with 24 numbered provisions, addressing AI-related disputes. The official text relies on the Civil Code, privacy, copyright, competition and procedural laws rather than creating a standalone AI statute.
The guidance covers unauthorized recognizable synthetic faces and voices, privacy, differentiated responsibilities for developers/providers/users, open-source licensing and AI-related evidence. Importantly, courts may collect evidence that parties cannot obtain themselves, consider preservation requests, and draw an adverse inference when a party unjustifiably withholds controlled evidence. That can matter when the model provider holds the relevant logs. It also addresses safeguards against fabricated AI evidence.
These are directions for adjudication, not measured litigation outcomes. The English explanatory article could not be opened directly; the primary Chinese text controls the account.
Why it matters
The concrete mechanism is an established court system applying existing rights and procedures to AI-specific information asymmetries. This fits more closely than a generalized claim that stronger AI regulation means dispersion. Potential access to evidence can help claimants, while rules protecting innovation, data interests and open-source activity distribute responsibility in other directions. The document does not demonstrate independent judicial oversight or successful remedies in practice.
