China sets a judicial benchmark for AI disputes

For privacy and artificial intelligence governance professionals, one persistent real-world question remains unsettled: When AI causes harm, including deep-fake impersonation, hallucinated defamatory statements, privacy breaches or copyright conflicts, who bears legal liability? 

Courts across jurisdictions have grappled with this question for years while Chinese courts have sometimes reached differing outcomes.

On 7 Sept., China's Supreme People's Court released its Opinions on Adjudicating Disputes Involving Artificial Intelligence in Accordance with the Law, setting a national judicial benchmark for AI-related disputes across the country. The opinions contain 24 provisions covering personality rights, personal information, intellectual property, consumer protection, autonomous vehicles and the use of AI in judicial proceedings. 

Previous Chinese AI rules have largely focused on how AI should be developed and used through administrative requirements such as algorithm filing, security assessments, content governance and labeling. The opinions serve a different purpose: promoting greater consistency in how courts across the country apply existing laws to AI disputes.

The opinions did not emerge overnight. In April, the SPC disclosed it was accelerating its drafting as Chinese courts confronted disputes involving AI-generated content, AI model parameters and data. Courts had concluded 908 cases involving data ownership and transactions in 2025 alone, a 25.6% increase from the previous year.

The legal form of the opinions is important. They were issued as a SPC judicial policy document rather than legislation or formal judicial interpretation. In China's legal system, such opinions do not create new law but guide courts in applying existing laws, including the Civil Code, Personal Information Protection Law and Copyright Law, to AI-related disputes. They carry significant practical authority. That choice matters as technology and legal questions evolve. China has not yet enacted a comprehensive AI law.

Fault liability becomes the general rule

One of the key features of the opinions is the SPC's approach to AI-related tort liability: Ordinary fault liability becomes the default for AI-related torts. The SPC explained that this approach is intended to avoid imposing excessive liability while AI technology and the industry remain at an early stage of development, which could discourage innovation.

China Civil Code recognizes ordinary fault liability, presumed-fault liability and strict, or no-fault, liability. Under ordinary fault liability, the plaintiff generally must establish the defendant's fault. Under presumed-fault liability, where specifically provided by law, the defendant bears the burden of showing an absence of fault. No-fault liability only applies where the law specifically provides for it.

The SPC makes ordinary fault liability under Article 1165(1) of China's Civil Code the general rule for AI-related torts unless existing law provides otherwise. In determining fault, courts should consider AI-specific factors, including the system's autonomy and technological transparency, potential risks, measures taken to prevent harm and the parties' ability to foresee and control those risks.

AI does not bear responsibility

The opinions do not treat AI itself as a responsible legal actor. Responsibility rests with the people or entities involved in developing, providing or using AI systems, depending on their conduct and applicable law. The fact that harmful content was generated through AI does not itself relieve them of responsibility.

The U.S. has no comparable nationwide tort standard for the AI industry; courts generally apply existing state doctrines case-by-case. In March, a California jury found Meta negligent in designing Instagram and the company liable for harm to a young user. China's approach differs because the SPC has made ordinary fault liability the default for AI-related torts nationwide.

From training data to AI-generated content

For AI model training, processing personal information that individuals have made public, or that has otherwise been lawfully made public, generally will not constitute infringement if the processing remains within a reasonable scope and the individual has not expressly objected. However, consent must be obtained, as required by law, where processing significantly affects an individual's rights and interests.

Courts determining what constitutes a reasonable scope should consider factors including the purpose of processing, its relationship to the model's functionality, the sensitivity of the information, potential effects on the individual and the uses the individual could reasonably have anticipated.

The opinions also address AI-generated content. If a generative AI provider receives proper notice that its service is generating content infringing another person's personality rights and fails to take necessary measures promptly, the provider may be liable for resulting harm. Users who deliberately use infringing prompts to induce harmful output may also be liable.

Intellectual property disputes receive particular attention. In copyright infringement cases, a developer asserting noninfringement may be required to produce evidence concerning training-data sources, training records and model operations. This is important where the provider controls the evidence needed to understand an AI system, giving courts a mechanism to address evidentiary problems created by opaque systems.

For companies, documentation of training-data provenance and model development may therefore become important not only for compliance but also for litigation. But documentation alone will not resolve the underlying copyright questions.

Significantly, the SPC left two consequential AI copyright questions unresolved: whether AI-generated content can itself qualify for copyright protection and how unauthorized use of copyrighted works for model training should be treated. The SPC acknowledged significant disagreement over these issues during drafting and concluded that further judicial experience is needed.

AI enters the courtroom

The opinions also address generative AI use by lawyers and litigants. Parties using AI to prepare court filings or case-research reports must verify the authenticity and accuracy of cited laws, judicial interpretations and cases, disclose their use of AI and remain responsible for submitted materials.

Using AI to fabricate evidence or create false litigation may result in fines, detention or, where the conduct constitutes a crime, criminal liability. Courts are also given guidance for evaluating AI-generated evidence. The principle is straightforward: AI may assist people in judicial proceedings, but it does not replace human responsibility.

A broader enforcement signal

The opinions make clear that AI misconduct may lead not only to civil remedies but also to fines, detention and, where conduct constitutes a crime, criminal liability. They address AI-fabricated evidence and false litigation, as well as AI-enabled fraud, defamation and unlawful acquisition or use of personal information.

The SPC also calls for coordination among courts, cyberspace regulators, public security authorities, procuratorates and market regulators, seeking a broader system combining industry self-regulation, administrative supervision and judicial protection.

For companies, the opinions therefore matter well before a dispute reaches court. How a company manages training data, implements safeguards, maintains technical records and responds to infringement complaints may later affect a court's determination of liability.

China's approach also offers global privacy and AI governance professionals a useful point of comparison. The EU has built much of its AI framework through comprehensive legislation while the U.S. relies on a combination of existing laws, state regulation and case-by-case litigation. China is now adding nationwide judicial guidance on how existing law should apply to AI disputes.

These differences matter for companies assessing cross-border AI liability. China's AI governance is entering a new stage: Regulators shape how AI is developed and used while the SPC provides a national judicial benchmark for determining responsibility when AI causes harm.