China has become the first major economy to establish a nationwide judicial standard for AI liability, setting clear rules for who bears responsibility when artificial intelligence causes harm. On September 7, China's Supreme People's Court released its Opinions on Adjudicating Disputes Involving Artificial Intelligence in Accordance with the Law, a 24-provision framework addressing everything from deepfake impersonation and defamatory statements to copyright conflicts and autonomous vehicle accidents.

The timing reflects a real problem courts have struggled with for years: when AI generates harmful content or violates someone's rights, who gets sued? Chinese courts had reached different conclusions on similar cases, creating legal uncertainty. Now, the Supreme People's Court has provided nationwide guidance on how existing laws should apply to AI disputes, making it easier for companies to understand their exposure and for courts to reach consistent decisions.

Why Is This Different From How the U.S. Handles AI Liability?

The United States has no comparable nationwide standard. Instead, American courts apply existing state-by-state doctrines case-by-case, creating a patchwork of liability rules. In March 2026, a California jury found Meta negligent in designing Instagram and held the company liable for harm to a young user, but that ruling applies only to that specific case. China's approach is fundamentally different: the Supreme People's Court has made ordinary fault liability the default standard for AI-related torts across the entire country.

The European Union, by contrast, has built its AI framework through comprehensive legislation like the AI Act. China is taking a middle path, using judicial guidance to clarify how existing laws apply to AI without waiting for new legislation. This matters because companies operating globally now face three distinct regulatory approaches, each with different liability implications.

What Does "Ordinary Fault Liability" Actually Mean for Companies?

Under China's Civil Code, there are three types of liability: ordinary fault, presumed fault, and strict liability. With ordinary fault, the person suing generally must prove the defendant was careless or negligent. The Supreme People's Court chose this standard as the default for AI-related torts because it avoids imposing excessive liability while AI technology remains in early stages, which could discourage innovation.

However, ordinary fault doesn't mean companies get a free pass. When determining whether a company was at fault, courts must consider AI-specific factors including the system's autonomy level, how transparent the technology is, potential risks, what safeguards the company implemented, and whether the company could reasonably foresee and control those risks.

How to Prepare Your AI Systems for Legal Scrutiny

The opinions make clear that documentation and responsible practices matter long before any dispute reaches court. Companies should focus on these key areas:

• Training Data Records: Maintain detailed documentation of where training data came from, how it was processed, and what safeguards were used. In copyright infringement cases, developers may be required to produce evidence about training-data sources and model operations, giving courts a way to understand opaque AI systems.
• Safeguard Implementation: Document what measures you took to prevent harmful outputs, including content filters, monitoring systems, and response protocols. Courts will evaluate whether your safeguards were reasonable given the risks your system posed.
• Rapid Response Procedures: If a generative AI provider receives notice that its service is generating infringing content and fails to take necessary measures promptly, the provider may be held liable. Having clear procedures to investigate and remove problematic content is now a legal necessity.
• User Responsibility Tracking: Users who deliberately use infringing prompts to induce harmful output may also be liable. Document how you inform users of their responsibilities and what happens when they misuse your system.

What About AI-Generated Content and Copyright?

The Supreme People's Court left two major copyright questions deliberately 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 court acknowledged significant disagreement over these issues during drafting and concluded that further judicial experience is needed before setting a national standard.

However, the opinions do address personal information used in 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 stays within a reasonable scope and the individual has not expressly objected. But consent must be obtained where processing significantly affects an individual's rights and interests.

What Happens When AI Is Used in Court?

The opinions specifically 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. They must also disclose their use of AI and remain responsible for all submitted materials.

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

Who Actually Bears Responsibility When AI Causes Harm?

The opinions make clear that AI itself is not treated 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 anyone of responsibility.

This principle extends beyond civil liability. The opinions address AI-enabled fraud, defamation, and unlawful acquisition or use of personal information. AI misconduct may lead not only to civil remedies but also to fines, detention, and criminal liability where conduct constitutes a crime.

What Does This Mean for Global Companies?

The Supreme People's Court 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. This coordination signals that AI governance in China will be comprehensive and multi-layered.

For companies assessing cross-border AI liability, these differences matter significantly. China's approach offers a useful point of comparison to the EU's comprehensive legislation and the U.S. reliance on existing laws and case-by-case litigation. 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, not just in China but potentially in how regulators worldwide view responsible AI practices.

The opinions were issued as a 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 because they provide the consistency Chinese courts had previously lacked.