What happened
On 7 September 2026, China's Supreme People's Court (SPC) formally released the 'Opinions on Lawfully Hearing Cases Involving AI-Related Disputes' (最高人民法院关于依法审理涉人工智能纠纷案件的意见, Fafa [2026] No. 10) at a televised press conference led by SPC Vice President Tao Kaiyuan. This is the SPC's first judicial guidance specifically addressing AI disputes and is binding on all levels of Chinese courts. It sets liability rules for: (1) non-consensual AI deepfakes/voice cloning as infringement of personality/voice rights; (2) algorithmic price discrimination against consumers; (3) 'AI hallucination' content generated by generative AI systems; (4) generative AI service-provider liability where providers fail to act after notice of infringing content; (5) user liability for deliberately prompting AI to generate infringing content; (6) liability for autonomous/driver-assistance vehicle traffic accidents; and (7) AI model training practices. It also creates an expedited injunction mechanism for cases of serious, irreversible harm (e.g., deepfaked defamatory sexual content) and calls for enhanced inter-agency coordination (with CAC, police, prosecutors, market regulators) and international judicial cooperation on cross-border AI/data disputes.
Why it matters
This is the first nationwide, binding judicial rule set in China addressing how courts must adjudicate AI-related civil liability — filling a gap given China has not yet enacted a comprehensive AI law. It directly shapes litigation risk and compliance obligations for any AI developer, platform, or deployer operating in China, including foreign firms, by clarifying when generative AI providers, downstream users, and businesses using algorithmic pricing face civil/legal liability. Combined with China's parallel CAC content-moderation crackdown (5.61M items removed, 49,000+ accounts penalized, reported Sept 3), this signals a coordinated multi-agency enforcement push on AI harms in the world's second-largest AI market.
Action needed
AI developers, generative AI platforms, and enterprises deploying algorithmic pricing or AI-generated content features in China should review consent/notice-and-takedown procedures for likeness/voice replication, audit algorithmic pricing practices for discriminatory patterns, and establish rapid response protocols for rights-infringement notices to avoid liability exposure under the new SPC framework.