On September 7, 2026, China's Supreme People's Court issued Fa Fa [2026] No. 10, Opinions on Lawfully Hearing Disputes Involving Artificial Intelligence. The 24-point document guides courts on AI-related tort, privacy, personal-information, consumer, product-liability, autonomous-driving, intellectual-property, contract, data and evidence disputes.

It generally applies fault-based liability unless law provides strict or presumed-fault liability. Generative-AI providers may incur liability if, after a rights-holder notice supported by preliminary evidence, they fail to take necessary measures against infringing output.

For AI products, courts must assess unreasonable danger by considering intended use, autonomous learning, updates, user control, compliance with national or industry standards, and whether limitations and foreseeable risks were truthfully explained and clearly warned.

The guidance also addresses automated-driving accident records, training on publicly available personal information, algorithmic price discrimination, open-source software responsibility, AI-assisted inventions, data rights and AI-generated court materials. It is judicial guidance rather than a product certification, spectrum, import or labeling rule.

Impact

Impact assessment: High.

Key reasons: the Supreme People's Court describes this as the first AI adjudication-rule document issued by China's highest judicial body, and it directly shapes how courts evaluate provider conduct, product defects, warnings, personal-information use, evidence and damages.

Who is affected?

AI developers and providers, manufacturers and sellers of AI-enabled products and vehicles, data controllers, open-source contributors, platform operators and litigants.

Conclusion:

companies operating in China should review notices and takedown controls, training-data governance, product-risk disclosures, event-data retention and litigation evidence practices against the 24 points.