On 7 September 2026, the Supreme People’s Court (SPC) issued the Opinions on Adjudicating Disputes Involving Artificial Intelligence in Accordance with Law (Fa Fa [2026] No. 10). The 24-provision document is the SPC’s first set of adjudication guidance dedicated to AI disputes. In the absence of a standalone AI law in China, the Opinions provide more specific guidance on how existing laws - including the Civil Code, Copyright Law, Patent Law, Anti-Unfair Competition Law and Personal Information Protection Law - should be applied to AI-related disputes.
For overseas companies, rights holders and applicants engaged in AI research and development, product deployment, content generation, open-source software or data-related business in China, the following IP issues are particularly relevant:
• Copyright liability for AI-generated content: Liability may be allocated among developers, service providers and users by reference to factors including the type of AI service, sources of training data, each party’s level of participation, preventive measures taken and profits obtained. A developer relying on a non-infringement defence may be required to provide information on training-data sources, training records, model operation and relevant technical basis.
• AI-assisted inventions: An AI-related invention may qualify for patent protection where it uses technical means to solve a technical problem and achieves a technical effect. A natural person who makes a creative contribution to the substantive features of an AI-assisted invention may be recognised as the inventor. Patent specifications must still satisfy the sufficiency-of-disclosure requirement.
• Open-source software: In assessing liability, courts will consider the type of open-source licence, restrictions on rights, compliance measures and the level of information disclosure. An open-source developer or provider that supplies code modules free of charge and discloses their functions and security risks may, in appropriate circumstances, be exempt from liability for infringement resulting from another party’s use.
• Data rights: Lawfully obtained data in which a party holds relevant rights and interests may be protected. Depending on the circumstances, data or datasets may receive protection under copyright, trade secret law or the Anti-Unfair Competition Law.
• Evidence and litigation: AI-generated content may be used as evidence of infringement, but courts will examine issues such as prompt design, consistency of repeated outputs, similarity to the asserted work, model training and content-filtering mechanisms. A party controlling key electronic evidence that refuses to produce it without legitimate reason may face an adverse inference.
The Opinions signal a more concrete phase in the judicial treatment of AI-related IP disputes in China. Overseas rights holders and businesses should consider reviewing training-data provenance, model-development records, open-source licence compliance, documentation of human inventive contribution and preservation of electronic evidence before launching AI products or building IP portfolios in China.
Source: Supreme People’s Court, 7 September 2026
https://www.court.gov.cn/zixun/xiangqing/511101.html
SPC: Opinions on Adjudicating Disputes Involving Artificial Intelligence in Accordance with Law
This is an unofficial English translation prepared for international readers. Functional English equivalents or brief explanations are used for certain China-specific institutional, policy and internet terms where helpful. In the event of any discrepancy, the Chinese original prevails.
SPC Document No. 10 of 2026 (Fa Fa [2026] No. 10)
For the purposes of fully implementing the guiding principles of the 20th National Congress of the Communist Party of China and subsequent plenary sessions of the 20th CPC Central Committee, supporting the accelerated development of new quality productive forces (innovation-driven productivity) through rigorous and impartial adjudication, and guiding courts at all levels in adjudicating disputes involving artificial intelligence in accordance with law, the following Opinions are issued pursuant to the Civil Code of the People’s Republic of China, the Cybersecurity Law of the People’s Republic of China, the Data Security Law of the People’s Republic of China, the Copyright Law of the People’s Republic of China, the Anti-Unfair Competition Law of the People’s Republic of China, the Law of the People’s Republic of China on the Protection of Consumer Rights and Interests, the Personal Information Protection Law of the People’s Republic of China, the Civil Procedure Law of the People’s Republic of China and other relevant laws, and in light of judicial practice.
I. General Requirements
(1) Guiding Approach
Guided by Xi Jinping Thought on Socialism with Chinese Characteristics for a New Era, courts shall implement Xi Jinping Thought on the Rule of Law and the policy principles concerning China’s development as a strong cyber nation. Recognising artificial intelligence as a major driver of the new wave of technological revolution and industrial transformation, courts shall give equal weight to development and security, combine support for innovation with law-based governance, and adjudicate AI-related disputes fairly and in accordance with law. Judicial work should help improve the safety, reliability, controllability and fairness of AI applications and provide effective judicial support for China’s modernisation, national development and rejuvenation.
(2) Basic Principles
Adopt a human-centred approach. Judicial decisions should guide the use of AI in ways that enhance public welfare and strengthen judicial protection of individual rights and interests in the digital age. Courts should prudently explore and improve adjudication rules and mechanisms for matters such as algorithmic ethics and algorithmic rules, make full use of existing laws, and appropriately address conflicts of rules, social risks and ethical challenges arising from AI development. Courts should also encourage stronger corporate social responsibility in the AI sector, prevent and eliminate algorithmic discrimination, protect individuals’ rights to information and choice in automated decision-making, and prevent and regulate misuse of AI that infringes lawful rights and interests, so as to promote responsible and beneficial AI development.
Support innovation and development. Courts should respect the nature of scientific and technological innovation and the practical development of the AI industry, and support technological and industrial innovation through decisions that are consistent with law, policy and technological development. New approaches to judicial protection should be explored for emerging rights and interests, with stronger protection for AI-related innovations, innovators, innovative activities and the innovation environment. Rights protection and technological and industrial development should be balanced in an inclusive and prudent manner, creating an environment that encourages experimentation and tolerates reasonable failure. Monopolistic conduct and unfair competition carried out through AI should be regulated in accordance with law, while developers and users of AI should enjoy equal rights, opportunities and rules. Particular attention should be given to equal protection for small and medium-sized enterprises and other market participants. Fair and efficient adjudication of AI-related disputes should support deeper integration of the internet, big data and AI with the real economy, foster the AI industry, and maximise AI’s potential to support high-quality economic and social development.
Maintain a firm safety baseline. Legal liability should be determined accurately in light of the specific harm that an AI application may cause to individuals or entities, as well as the nature and scale of the relevant risk. In allocating liability among developers, providers, users and other parties, courts should distinguish between different types of large models, including general-purpose and specialised models and open-source and closed-source models, taking into account differences in technical architecture, risk spillover and capacity for control. The criminal policy of combining leniency with severity should be applied accurately. Criminal conduct that deliberately uses AI to endanger national security, infringe individual rights or disrupt public order shall be punished strictly in accordance with law, while innovative conduct in AI research, development and application shall be handled with appropriate caution and in accordance with the principles of legality and proportionality in criminal punishment.
II. Proper Adjudication of AI-Related Tort Cases and Protection of Lawful Rights and Interests
(3) Principles for Determining Liability in AI-Related Torts
Relevant provisions of the Civil Code, the Personal Information Protection Law, the Product Quality Law, the Road Traffic Safety Law and other laws shall be applied accurately when determining liability arising from the use of AI to infringe civil rights and interests. Where the law does not expressly provide for strict liability or a presumption of fault, liability shall be determined under the fault-based principle set out in Article 1165(1) of the Civil Code. In assessing whether a party was at fault and the degree of fault, courts should consider the particular application scenario, degree of autonomy, technical and informational transparency, potential risks and scope of impact; the measures taken by developers, providers and other relevant parties to prevent or reduce infringement and the technical feasibility of such measures; and the user’s ability to foresee and control the harm that may result from infringing conduct carried out using AI, among other factors.
(4) Generative AI Infringement of Personality Rights, Including Face-Swapping, Voice Cloning and “AI Resurrection”
Use of generative AI to process the name, image or other identifying attributes of a specific living or deceased individual must comply with laws and regulations and must not violate public order or good morals. Unless otherwise provided by law, where a person’s name, likeness or other identifying information is processed without consent to generate, use or publish an identifiable virtual digital representation of that person, the court shall support claims for infringement of the person’s rights to name, likeness or other personality rights. Unless otherwise provided by law, where a person’s voice is used without consent as training material to imitate the person’s timbre, intonation, pronunciation style or other vocal characteristics and generate an identifiable synthetic voice, the court shall support a claim that the person’s rights and interests in their voice have been infringed. Where an identifiable synthetic image or voice is manipulated to carry out improper conduct or make false statements that lower the social reputation of the individual concerned or another person, the court shall find an infringement of the right to reputation. Where AI is used without authorisation to create or use a virtual digital representation of a deceased person, thereby infringing the deceased person’s name, likeness, reputation or related interests, the court shall support a claim by the deceased person’s close relatives for civil liability under Article 994 of the Civil Code.
(5) AI-Enabled Doxxing and Other Infringements of Privacy
Where AI is used for the purpose of intruding into a person’s privacy by tracking or analysing publicly available information such as telephone numbers, online accounts or social-media activity in order to obtain private information, disclose or publish such information, or disturb the person’s private life, the conduct shall be regarded as an infringement of privacy. This includes AI-enabled doxxing and crowdsourced online identification sometimes referred to in China as “human flesh search”. Using AI to photograph, observe, monitor or listen in on a person’s private space or private activities shall also be regarded as an infringement of privacy, unless otherwise provided by law or expressly consented to by the individual.
(6) Civil Liability for Use of Personal Information in AI Training
For AI model training, processing within a reasonable scope of personal information that an individual has made public, or that has otherwise been lawfully made public, will generally not be regarded as infringing personal-information rights where the individual has not expressly objected. Consent must be obtained in accordance with law where the processing has a significant impact on the individual’s rights and interests. In determining what constitutes a reasonable scope, courts should consider the purpose of the processing and whether it is necessary and appropriate for the model’s functions, the type and sensitivity of the personal information and its potential impact on the individual, the context in which the information was made public, and the scope of use that the individual could reasonably have expected.
(7) Liability of Generative AI Service Providers
Where content automatically generated by generative AI infringes another person’s reputation, privacy or other personality rights, and the service provider fails, after receiving notice from the rights holder, to take timely and necessary measures such as preventing further generation of the infringing content, the provider shall bear tort liability for the resulting harm in accordance with law. The notice should include prima facie evidence of infringement and the rights holder’s true identity information. Where an online user maliciously induces a generative AI system to produce infringing content, including by entering infringing prompts, and thereby causes harm, the user shall bear liability in accordance with law. If, after notice from the rights holder, the service provider fails to take timely and necessary measures such as stopping generation of the infringing content or blocking the relevant generation instructions, the court shall support claims against the user and the provider under Article 1195 of the Civil Code.
(8) Injunctions to Protect Personality Rights
Where an individual, legal person or unincorporated organisation has evidence that another party is using, or is about to use, AI to unlawfully infringe its personality rights and failure to stop the conduct promptly would cause irreparable harm, it may apply to the court for an injunction ordering the conduct to cease or requiring a network service provider or generative AI service provider to stop providing the relevant service. In granting such an injunction, the court shall consider the type of personality right involved, the manner of the unlawful conduct, and the likely scope and seriousness of harm, and shall ensure that the injunction does not go beyond what is necessary.
(9) Product Liability for AI-Enabled Products
Courts shall apply the definition of “product” under the Product Quality Law when determining whether an AI-enabled product embodied in a physical product is subject to product-liability rules. Where a defective AI-enabled product causes harm, the manufacturer and seller shall bear product liability in accordance with law. In determining whether such a product presents an unreasonable danger to personal or property safety, courts should consider its nature and intended use, self-learning capability, upgrades and updates, the degree of user control over the system, and compliance with applicable national or industry standards. Particular attention should be paid to whether the manufacturer and seller provided truthful explanations and clear warnings concerning the product’s appropriate use cases, inherent limitations and foreseeable risks.
(10) Algorithmic Price Discrimination and AI Celebrity Impersonation in Sales
Where a business uses algorithms to impose unreasonable differential treatment in price or other transaction terms for the same goods or services, thereby infringing lawful rights and causing harm, it shall bear corresponding tort liability in accordance with law. In determining whether differential treatment is unreasonable, courts should consider whether it materially restricts or harms consumers’ rights to information, independent choice or fair dealing; whether personalised transaction terms are generated from information such as purchasing preferences, willingness or ability to pay, or browsing history; whether the practice is inconsistent with good faith or business ethics; and whether the reasons for the differential treatment are legitimate, sufficient and non-discriminatory. Where AI is used to impersonate a celebrity in the promotion or sale of goods or services and the conduct constitutes fraud, the court shall support a consumer’s claim for punitive damages under Article 55 of the Consumer Rights Protection Law.
(11) Liability for Accidents Involving Autonomous or Driver-Assistance Vehicles
Where an autonomous vehicle or a vehicle equipped with driver-assistance functions is involved in a road accident that causes damage, liability for compensation shall be determined under the Civil Code and the Road Traffic Safety Law. If a product defect in the vehicle caused the accident, the court shall support a claim that the manufacturer or seller bear liability under Chapter IV of Book Seven of the Civil Code. Where a defect in a driver-assistance vehicle and the driver’s fault combine to cause the same harm, the court shall support claims against both the driver and the manufacturer or seller in accordance with Article 1172 of the Civil Code and other applicable provisions. If a manufacturer or seller makes false or misleading representations about the level of automation, intelligence, performance or intended use of an autonomous or driver-assistance vehicle and thereby harms consumer rights, the court shall support claims for civil liability under the Civil Code, the Consumer Rights Protection Law and other applicable laws. To determine the cause of an accident, a court may require a manufacturer, seller, operator or other party controlling relevant data to provide, to the extent necessary, true and complete autonomous-driving or driver-assistance event records and other data needed to establish the facts.
III. Proper Adjudication of AI-Related Intellectual Property Disputes
(12) AI-Related Infringement of Intellectual Property Rights
EN Where AI-generated content infringes another party’s copyright, courts shall determine the respective liability of AI developers, providers and users in a reasonable manner, taking into account the type of AI service, characteristics of the relevant industry, sources of training data, the degree of participation of each party, necessary measures taken to prevent infringement, and any resulting profits. Where an AI developer raises a non-infringement defence, the court shall require the developer to substantiate that defence by providing information such as the sources of training data, records of the training process, the model’s operating mode and relevant scientific or technical basis. A rights holder alleging that an AI provider used algorithmic technology to infringe copyright shall provide corresponding evidence. Where an AI user knew or should have known of a pre-existing work, uses AI to generate a work substantially similar to that earlier work, and has no reasonable defence, the court shall support the prior rights holder’s infringement claim. The use of AI for counterfeiting or other infringement, false advertising, or manipulation of traffic, reviews or transactions that infringes another party’s rights or constitutes unfair competition shall give rise to corresponding liability in accordance with law.
(13) Liability Relating to AI Open-Source Software
In cases involving open-source software, courts shall determine the infringement liability of open-source developers and providers, as well as downstream developers and providers, by considering factors including the type of open-source licence, the specific restrictions on rights, security and compliance measures, and the extent of information disclosure. Appropriate limitations or exemptions from liability may be afforded to open-source developers and providers in accordance with law. Where a developer or provider makes certain code modules required for AI software development available free of charge on an open-source basis, and publicly explains their functions and security risks, the court may find that the developer or provider is not liable for infringement caused by another party’s use of those modules.
(14) Patentability, Inventorship and Disclosure for AI-Related Inventions
An AI-related invention shall be regarded as patent-eligible subject matter where it uses technical means consistent with the laws of nature to solve a technical problem and achieves a technical effect consistent with the laws of nature, except where it violates law or public morality, harms the public interest, or no natural person has made a substantive contribution. Where a natural person uses AI in making an invention and makes a creative contribution to the substantive features of the invention, that person shall be recognised as the inventor. An AI-related patent specification shall satisfy the sufficiency-of-disclosure requirement where its description of the technical solution enables a person skilled in the art to carry out the invention.
(15) Contractual Liability in AI Technology Agreements
When adjudicating disputes arising from AI-related technology development, transfer, licensing, consultancy or service agreements, courts shall determine contractual liability in accordance with the agreement and the law, taking full account of the characteristics of AI research and development and whether the technology developer has exercised reasonable efforts, among other circumstances.
(16) Use and Protection of Data in AI Activities
Where an AI developer lawfully obtains data through collection and generation, derivative creation, assignment, licensing or other lawful means and holds corresponding rights and interests in that data, the court shall protect those rights and interests. Data or datasets that constitute compilation works or otherwise meet the requirements for copyright protection shall be protected under the Copyright Law. Data or datasets that constitute trade secrets shall be protected under the Anti-Unfair Competition Law. Where data or datasets do not constitute trade secrets, liability may nevertheless arise where the challenged conduct violates Article 13 of the Anti-Unfair Competition Law. Businesses that use data, algorithms or other technological means to enter into monopoly agreements or abuse a dominant market position shall bear corresponding liability in accordance with law. Liability shall also arise where technical means such as fabricated disruptive data, malicious data labelling or adversarial-example attacks are used to compromise the safe operation of AI systems.
IV. Procedural Rules for AI-Related Disputes and Misuse of AI in Judicial Proceedings
(17) Fact-Finding in AI-Related Disputes
Given the technical and specialised nature of AI-related disputes, courts shall provide appropriate case-management guidance and judicial clarification to ensure that the parties present evidence fully and promptly on the disputed facts. A party that cannot collect evidence itself for objective reasons may apply to the court for investigation and collection; where necessary, the court may also collect evidence on its own initiative. Applications for preservation of evidence concerning key technologies shall be examined promptly. If a party in possession or control of documentary evidence, electronic data or other evidence refuses to produce it without a legitimate reason, the court may accept the opposing party’s assertion that the evidence would have been adverse to the party withholding it. In cases involving specialised issues such as AI technical principles or operating mechanisms, courts should make appropriate use of lay assessors, forensic experts, expert assistants and technical investigation officers to assist in establishing the facts.
(18) Review and Assessment of Evidence in AI-Related Disputes
Courts should continue to develop evidentiary review rules suited to new technologies, taking full account of the complexity of AI, the opacity of system operation and the particular difficulties of evidence collection. Particular scrutiny should be given to the authenticity and integrity of electronic data throughout its generation, collection, storage and transmission. Evidence produced through different technologies, including big data and blockchain, should be assessed according to technology-specific considerations. For big-data analysis reports, courts should focus on matters such as the source of the original data, data-cleaning rules and the scientific reliability of the analytical method. For blockchain-based evidence preservation, courts should focus on the authenticity of the data before it was recorded on the blockchain and the reliability of the technical platform. Where AI-generated content is submitted as evidence of infringement, courts should consider factors including the design of the prompts and their effect on the output, the degree of similarity between the generated content and the work in which rights are asserted, the consistency of repeated testing, and relevant aspects of model training, algorithm design and content-filtering mechanisms.
(19) Improper Use of AI to Create Evidence or Disrupt Judicial Proceedings
Where a party exploits AI’s self-learning or autonomous decision-making capabilities to create false evidence through human intervention or misleading techniques - such as deleting or altering labels identifying generated or synthetic content, using specific instructions, selectively presenting results, or applying adversarial interference - and fabricates material facts in order to bring sham civil proceedings and harm state interests, the public interest or another party’s lawful rights, the court shall dismiss the claims and may impose a fine or judicial detention depending on the seriousness of the conduct. Criminal liability shall be pursued where the conduct constitutes a crime. A participant in proceedings or any other person who uses AI to forge evidence and obstruct the court’s handling of a case shall be dealt with under Article 114 of the Civil Procedure Law. Where pleadings, case-law research reports or other materials submitted in proceedings have been generated using AI, the submitting party must carefully verify the authenticity and accuracy of the cited laws, judicial interpretations, cases and other content before filing, disclose the use of AI assistance when submitting the materials, and bear legal responsibility for their authenticity and accuracy.
(20) Criminal Conduct Committed Using AI
Where the use of AI to commit fraud, insult or defamation, damage business or product reputation, infringe personal information, unlawfully obtain computer information-system data, or produce, sell or disseminate obscene materials constitutes a crime, criminal liability shall be pursued in accordance with law. Criminal liability shall also be pursued where, after activating a driver-assistance function, a person uses unauthorised accessories to evade monitoring by the driver-assistance system and thereby causes a road accident, if the conduct constitutes a crime.
V. Judicial Guidance and Supporting Mechanisms for AI-Related Disputes
(21) Use of Multiple Dispute-Resolution Mechanisms
Courts should actively use multiple dispute-resolution mechanisms to resolve AI-related disputes appropriately, including by coordinating with competent AI industry authorities, specialised mediation organisations, AI experts and scholars, and by making effective use of pre-litigation mediation. Through their adjudicative role, courts should strengthen law-based prevention and resolution of AI disputes, clarify rights, obligations and boundaries of conduct through open hearings and case-based legal guidance, and use judicial decisions to support early prevention and timely resolution of disputes.
(22) Strengthening Judicial Supervision and Guidance
The mechanism for transferring appropriate cases to a higher-level court for trial shall be applied in a standardised manner to strengthen judicial supervision and guidance and the precedential value of significant cases. AI-related disputes involving major interests, novel and complex issues, rule-setting significance, or a need to unify standards for the application of law may be tried by a higher-level court. Courts shall make full use of the People’s Court Case Database and strengthen case guidance in order to promote consistent application of law and further development of AI governance rules.
(23) Inter-Agency Coordination
Courts shall make full use of their judicial functions, including through judicial recommendations and cooperation mechanisms, to support effective governance. Coordination shall be strengthened with cyberspace regulators, the police force, prosecutorial authorities, market regulation authorities and other relevant agencies, with long-term mechanisms for coordinated governance to be established and implemented. Relevant businesses, platforms and industry associations should be encouraged to strengthen self-governance, so as to create an integrated framework combining industry self-regulation, administrative oversight and judicial protection across the AI sector.
(24) International Judicial Exchange and Cooperation
Courts shall properly adjudicate disputes involving AI and cross-border data. On the basis of mutual respect, equality and mutual benefit, courts shall promote exchange and cooperation in the adjudication of foreign-related AI cases, strengthen alignment and mutual learning of relevant rules, and support a fairer, more reasonable, inclusive and open AI governance framework that promotes and protects the diversity of AI ecosystems and human civilisation.
Supreme People’s Court
7 September 2026
Source：https://www.court.gov.cn/zixun/xiangqing/511101.html
