China Unveils New Rules for Settling AI Disputes and Deepfake Conflicts

China has released a new set of rules detailing how courts, regulators and companies should handle disputes involving artificial intelligence — from faulty generative‑AI outputs to deepfakes that damage a person’s reputation. The provisions place clear obligations on AI service providers to spell out terms of use, maintain complaint‑handling systems, and keep records that make it possible to trace which model generated a particular piece of content.

Crucially, the rules push back against the idea that platforms can shrug and say “the AI did it.” When users suffer harm, providers are expected to investigate, explain what went wrong, and take corrective steps; serious cases of malicious synthetic media can trigger penalties. Taken together, the package is meant to serve as a reference point for Chinese courts and administrative agencies as AI‑related conflicts become more common.

While the US and EU are still arguing over broad, omnibus AI acts, Beijing is steadily building a patchwork of highly specific rules targeted at concrete use cases. For policymakers in countries like Japan, where deepfake abuse and AI‑driven copyright infringement are already emerging, China’s “dispute‑resolution first” approach offers a very different model of governance to study — even if its enforcement context is not directly transferable.