The Opinions of the Supreme People’s Court on Lawfully Trying Cases Involving Artificial Intelligence Disputes were issued and took effect on September 7, 2026.
Artificial intelligence has developed at an extremely rapid pace and permeated all aspects of people’s work and daily life. Consequently, the number of AI-related cases continues to rise, presenting diverse fact patterns. To address this, the Supreme People’s Court issued and implemented the Opinions of the Supreme People’s Court on Lawfully Trying Cases Involving Artificial Intelligence Disputes (Fa Fa〔2026〕No.10） on September 7, 2026.
Below is a brief analysis and explanation of the key provisions of the Opinions:
- Imputation Principles for Tort Liability
Pursuant to Article 1165 of the Civil Code, the prerequisite for establishing tort liability is proof that the actor is at fault, unless the law explicitly prescribes a reversal of the burden of proof. In AI torts, it is often difficult to determine whether the actor is at fault for consequences arising from AI out-of-control incidents. Subparagraph （3） of Article 2 of Fa Fa〔2026〕No.10 stipulates: “…When determining whether an actor is at fault and the degree of such fault, the people’s court shall comprehensively consider factors including the specific scenario of the AI application, its level of autonomy, technical and informational transparency, potential risks and scope of impact; measures adopted and technical feasibility for AI developers, providers and other relevant parties to prevent and mitigate AI-related torts; and the foreseeability and control capacity of AI users over harm that may result from torts committed using AI technologies.”
Accordingly, for manufacturing enterprises deploying AI robotic arms and robots for operations, key factors to be assessed in future disputes shall include: whether the enterprise possesses foreseeability and control capacity over potential harm that such equipment may inflict on other participants in the operation, operators and other related personnel; and whether the enterprise has implemented effective preventive and control measures.
- AI Infringement of Copyright
Over the past few years, there have been divergent views on whether information generated by AI constitutes copyright infringement. Fa Fa〔2026〕No.10 does not provide a universal conclusion, but sets out factors courts shall consider when hearing such infringement cases and specifies the burden of proof borne by each party. Subparagraph （12） of Article 3 of Fa Fa〔2026〕No.10 provides: “…People’s courts shall comprehensively consider factors such as the type of AI service, industry characteristics, sources of training data, degree of participation of each party, necessary measures adopted and profit gains, to lawfully and reasonably determine the liabilities of AI developers, providers and users. Where an AI developer raises a non-infringement defense, the people’s court shall order the developer to provide supporting evidence including sources of training data, records of the training process, model operation modes and scientific theoretical basis. Where a right holder claims that an AI provider infringes its copyright using algorithmic technology, the right holder shall submit relevant evidence. …”
- Patent Grant and Validation for Inventions Created Using AI
Is AI merely a tool, or a secondary human brain capable of creativity? Fa Fa〔2026〕No.10 leans toward the former. Subparagraph (14) of Article 3 stipulates: “Where an AI-related invention creation adopts technical means complying with natural laws, solves a technical problem and achieves technical effects in line with natural laws, people’s courts shall lawfully recognize it as subject matter protected by the Patent Law, except where such invention creation violates the law, social morality, harms public interests, or involves no substantive contribution by a natural person. …”
Therefore, in future patent disputes concerning AI-generated inventions, whether a natural person has made substantive contributions will become a critical point of contention. Where enterprise R&D departments utilize AI, it is recommended that they preserve evidence of substantive R&D input by researchers, particularly records of ideas and research directions. In addition, costs incurred from AI use may be counted toward R&D expenditure, which will be helpful to calculate damages in the event of trade secret or patent infringement litigation.
Fa Fa〔2026〕No.10 also addresses more than ten other dispute points including AI data usage, liability for breach of technical contracts, and infringement of consumer rights and interests, which merit attention.