The head of the relevant department of the Supreme People's Court answered questions from reporters on the "Opinions on the Trial of Disputes Involving Artificial Intelligence in Accordance with the Law"

On September 7, 2026, the Supreme People's Court held a press conference on the "Opinions of the Supreme People's Court on the Trial of Disputes Involving Artificial Intelligence in Accordance with the Law". Tao Kaiyuan, Vice President of the Supreme People's Court, Zhou Jiahai, Director of the Research Office, Li Jian, President of the Third Civil Court, and Si Yanli, Deputy Director of the Research Office, attended the press conference and answered questions from reporters. The press conference was presided over by Ji Zhongbiao, the spokesperson of the Supreme People's Court.

Q: We have noticed that the "Opinions of the Supreme People's Court on the Trial of Disputes Involving Artificial Intelligence in Accordance with the Law" (hereinafter referred to as the "Opinions") are based on current hot and difficult issues, and provide targeted responses to the concerns of the people and the artificial intelligence industry. May I ask, in terms of specific content arrangement, how does the Opinion properly handle the relationship between rights protection and innovative development?

Answer: General Secretary Xi Jinping pointed out in his keynote speech at the opening ceremony of the 2026 World Conference on Artificial Intelligence and the High level Conference on Global Governance of Artificial Intelligence that we must always adhere to the equal importance of development and security, grasp the trends and laws of artificial intelligence development, continuously improve relevant laws, regulations, policy systems, application norms, and ethical standards, ensure the safety, reliability, and controllability of artificial intelligence, and make it run fast and steadily as a thousand mile horse. The Opinion is closely centered around the principle of "balancing development and security". It not only focuses on regulating the illegal and disorderly use of artificial intelligence in accordance with the law, strengthening judicial protection of people's rights and interests in the digital age, but also clearly supports technological and industrial innovation, and safeguards the healthy development of artificial intelligence. Specifically:

In terms of safeguarding people's livelihood rights and interests, the "Opinions" mainly make the following provisions: firstly, increase the judicial protection of personality rights and interests. Currently, illegal activities such as AI face swapping, AI voice imitation, and "network box opening" occur frequently, which is of great concern to all sectors of society. To this end, the "Opinions" have made clear regulations on the accountability for the use of artificial intelligence to infringe on portrait rights, reputation rights, privacy rights, voice rights, and other behaviors, in order to fully play the role of judicial judgment norms and guidance, and promote the improvement of artificial intelligence. Secondly, protect personal information rights and interests in accordance with the law. Large scale and high-quality training data are the cornerstone of the development of artificial intelligence technology. In order to balance and balance the protection of personal information rights and interests with the promotion of the development of the artificial intelligence industry, the "Opinions" are based on legal provisions such as the Civil Code and the Personal Information Protection Law. It is explicitly stated that training artificial intelligence models to process legally disclosed personal information within a reasonable range, and individuals who do not explicitly refuse, is generally not considered infringement. However, if it has a significant impact on individual rights and interests, personal consent should be obtained in accordance with legal provisions. Thirdly, regulate in accordance with the law behaviors that infringe upon the legitimate rights and interests of consumers, such as "big data murder" and "counterfeit celebrity sales". Recently, the "big data murder" triggered by AI algorithms has been repeatedly exposed, and the "counterfeit celebrity sales" have also attracted much attention. We cannot expect consumers to have sharp eyes. The law must be promptly enforced to safeguard the legitimate rights and interests of consumers. For this reason, the Opinion clarifies that if an operator uses algorithms to implement unreasonable differential treatment in transaction prices and other transaction conditions for the same product or service, causing harm to the legitimate rights and interests of others, the people's court shall determine that they shall bear corresponding liability for infringement in accordance with the law; If a consumer claims punitive damages for using artificial intelligence to "counterfeit celebrity sales" and constitutes fraud, the people's court shall support it in accordance with the law.

In terms of supporting innovative development, the "Opinions" mainly make the following provisions: firstly, accurately grasp the principle of attribution of liability for artificial intelligence infringement, and avoid bearing excessive responsibility at the early stage of the development of artificial intelligence technology and industry, which may affect the enthusiasm for innovation. The Opinion clarifies that the legal liability arising from the use of artificial intelligence to infringe upon civil rights and interests shall be subject to the principle of fault liability, except where the law explicitly provides for the application of no fault liability or presumed fault liability. Secondly, strictly recognize "artificial intelligence products" in accordance with the law to avoid the generalization of product liability. The Opinions clearly stated that AI products should be limited to products with physical carrier, such as intelligent robots, autonomous vehicle, etc., and AI services without physical carrier should be excluded in strict accordance with the definition of "product" in the Product Quality Law. Thirdly, comprehensively and accurately implement the criminal policy of balancing leniency and severity, and handle criminal cases involving artificial intelligence in accordance with the law. On the one hand, the Opinion emphasizes that criminal acts such as intentionally using artificial intelligence to endanger national security, infringe on citizens' rights and interests, and disrupt social order shall be severely punished in accordance with the law; On the other hand, it is required that innovative behaviors in the research and application of artificial intelligence be handled prudently in accordance with the law, and adhere to the principles of legality and proportionality between crime and punishment. In addition, as Vice President Tao introduced earlier, the "Opinions" have made special provisions for the proper trial of cases involving artificial intelligence intellectual property disputes in accordance with the law, and the relevant provisions fully reflect the concept of supporting innovative development on the track of the rule of law.

Q: The frequent occurrence of new types of intellectual property disputes caused by artificial intelligence generated content has attracted widespread attention. May I ask, what are the new provisions in the Opinion regarding the handling of intellectual property disputes involving content generated by artificial intelligence?

Answer: Articles 12 to 16 of the Opinion are specific provisions for intellectual property disputes involving artificial intelligence, among which Article 12 provides guidance on the issue of content generated by artificial intelligence that you mentioned. The outline of the 15th Five Year Plan clearly proposes to explore the establishment of rules for the ownership of rights of artificial intelligence generated products and the recognition of rights and responsibilities of developers, operators, and users. The recently released "15th Five Year Plan for Intellectual Property Protection and Utilization" by the State Council further emphasizes the need to improve the copyright system that adapts to the development of artificial intelligence. The Opinion implements the relevant arrangements of the central government and responds promptly to social concerns. There are three specific aspects to the situation:

One is to clarify the considerations for determining the responsibility of relevant parties in cases of copyright infringement caused by artificial intelligence generated content. The infringement of content generated by artificial intelligence involves multiple parties such as developers, providers, and users of artificial intelligence technology. If the content generated by artificial intelligence is accused of infringing on the copyright of others, the relevant parties should not be exempted from liability based on the fact that the accused infringing content is generated by artificial intelligence, but should adhere to the principle of matching responsibility with control ability and duty of care. Regarding the determination of control ability and duty of care, it is generally necessary to comprehensively consider the technical and business model characteristics of the involved artificial intelligence, the role played by the parties in the content generation process, who "feeds" the relevant training data, necessary measures taken, profit situation, and other factors. On this basis, the subjective fault and degree of fault of the parties involved in the infringement shall be determined in accordance with the law, and the liability for infringement shall be reasonably divided.

The second is to clarify the evidentiary rules for copyright infringement disputes arising from content generated by artificial intelligence. If the rights holder claims that the developer or provider of artificial intelligence services has infringed upon their copyright, they shall provide preliminary evidence of the fact that the infringing content was generated by the artificial intelligence and is substantially similar to their copyrighted works. Due to the complexity and opacity of artificial intelligence technology, relevant evidence cannot be obtained by others. Therefore, if artificial intelligence developers raise a non infringement defense, they should provide factual evidence such as the source of training data, training process, and operating mode of the model, and if necessary, provide relevant scientific theoretical basis for support. If an artificial intelligence user knows or should know about a prior work and uses artificial intelligence to generate a work that is substantially similar to the prior work, and there are no grounds for exemption such as reasonable use, it shall be deemed to constitute infringement.

The third is to regulate the use of artificial intelligence to generate content and engage in unfair competition in accordance with the law. In the past two years, disputes over unfair competition using artificial intelligence have been increasing. Operators use artificial intelligence generated graphics, videos, virtual character images, etc. to fabricate traffic and positive reviews, and even use AI face swapping to fabricate popular science videos for false advertising. These behaviors not only harm the legitimate rights and interests of others, but also seriously mislead consumers and disrupt market competition order. They should be legally recognized as constituting counterfeiting infringement, false advertising, etc.

It should be noted that during the drafting and argumentation process of the Opinion, there were significant differences in opinions regarding the copyrightability of content generated by artificial intelligence, as well as the qualitative issues of using others' works without permission to train artificial intelligence models. Further understanding is needed. Therefore, the Opinion does not make provisions on these two issues for the time being. Currently, artificial intelligence technology is developing rapidly, and new situations and problems are constantly emerging. Next, we will continue to closely monitor intellectual property protection issues in emerging fields such as artificial intelligence, continuously understand and grasp the rules in practice, constantly summarize trial experience, improve judicial rules, and strive to promote consensus.

Q: With the accelerated iteration of artificial intelligence, the threshold for infringement of technologies such as deep forgery, intelligent collection, and user profiling has been significantly reduced, and the protection of personal rights and interests is facing severe challenges. As mentioned earlier, the "Opinions" have made provisions on hot issues such as "AI face swapping and voice acting" and "network box opening" that infringe on personal rights and interests. Could you please provide more specific information?

Answer: Personality rights are related to everyone's personal dignity and are the most basic and important rights of civil subjects. We emphasize the need to balance the protection of personality rights and the development of the artificial intelligence industry with a tolerant and cautious attitude. However, tolerance does not mean indulgence, and prudence does not mean laissez faire. In response to social concerns and to guide judicial practice, the "Opinions" make targeted provisions for different types of behaviors that abuse and misuse artificial intelligence technology to infringe on personal rights and interests. Mainly includes:

One is to regulate in accordance with the law the use of artificial intelligence, such as "AI face swapping and voice mimicking", to infringe upon the rights of portrait, reputation, and voice. In practice, the most strongly reflected is AI deep forgery. For example, AI face swapping technology can indiscriminately "harvest" everyone's portrait, and ordinary people may also become targets for deep forgery; For example, using AI onomatopoeia technology, one can "steal sound" for a few yuan, and the effect is enough to be realistic. Recently, the news media also uncovered a black industry chain that sells "AI one click undressing" software and tutorials, which the people deeply condemn. Article 4 of the Opinion clearly stipulates that the use of artificial intelligence to process a natural person's name, portrait, etc. without consent, to generate a virtual digital image that can identify the natural person and use it publicly, constitutes an infringement of personality rights such as the right to name and portrait; Using natural person voices as training materials without consent, imitating their timbre, intonation, and pronunciation style to generate recognizable synthesized human voices, constitutes a violation of sound rights; Those who manipulate virtual images, synthesize sounds, engage in inappropriate behavior, make false statements, or lower the social evaluation of others constitute infringement of reputation rights, and the perpetrator shall be held liable for infringement in accordance with the law. In addition, in response to the phenomenon of "AI resurrecting the deceased" that has attracted social attention, the "Opinions" have also made specific provisions to protect the personal interests of the deceased in accordance with the law.

The second is to regulate in accordance with the law the use of artificial intelligence to carry out behaviors that infringe on the privacy rights of others, such as "network unboxing" and "human flesh search". 'Online unboxing' and 'human flesh search' are prominent forms of online violence, which not only seriously infringe on the privacy rights of victims and disturb their peace of life, but also affect the public's sense of security and pollute the online ecosystem. Currently, there is a phenomenon of using artificial intelligence technology to track and analyze public information such as phone numbers, online accounts, and social media of specific natural persons, to obtain private information, or to leak and disclose the obtained private information. In this regard, Article 5 of the Opinions has specifically made provisions.

The third is to clarify the applicable rules of the prohibition system for infringement of personality rights. The use of artificial intelligence to infringe upon personal rights and interests has the characteristics of low cost, fast dissemination, wide impact, and irreversible damage. The traditional remedy model of filing a lawsuit after the fact is not conducive to timely and effective protection of the legitimate rights and interests of victims. Therefore, the legal application of personality rights infringement injunctions is particularly important. Therefore, Article 8 of the Opinion clarifies this. For example, if the victim has evidence to prove that they have been falsely accused by AI and their legitimate rights and interests will be irreversibly damaged if not stopped in a timely manner, they can apply to the people's court for a personality rights infringement injunction. After examination, if the conditions are met, the people's court may order the network user to stop the relevant infringement behavior in accordance with the law, or order the network service provider or artificial intelligence service provider to take necessary measures in a timely manner to prevent further expansion of the damage.

It should be noted that currently, there are frequent occurrences of fraud, insult, defamation, infringement of citizens' personal information, production, sale, and dissemination of obscene materials using techniques such as deepfakes. If the relevant behavior constitutes a crime, it must be severely punished in accordance with the law and criminal responsibility shall be pursued. In this regard, the "Opinions" have made special provisions.

Q: Currently, generative artificial intelligence software is widely used, and related civil disputes have also attracted social attention. I have noticed that the Opinion has introduced the "safe harbor rule" for the infringement determination of generative artificial intelligence. Can you provide an in-depth introduction to this issue?

Answer: Article 1195 of the Civil Code stipulates the "Notice of Deletion Rule", commonly known as the "Safe Harbor Rule". Simply put, it refers to a network user using online services to commit infringement (such as posting insults to others online). After the rights holder discovers this, they have the right to notify the network service provider to take measures such as blocking or deleting. After receiving a qualified notification, the network service provider should take necessary measures in a timely manner. Otherwise, they will be jointly and severally liable with the network user for the expanded damage. On the contrary, if the network service provider receives the notification and takes necessary measures in a timely manner, they usually do not need to bear infringement liability.

In practice, there is controversy over whether providers of generative artificial intelligence services can advocate for the application of the "safe harbor rules". We believe through research that although generative artificial intelligence services differ from traditional network services, there are also similarities in their operational mechanisms. For example, generative artificial intelligence service providers also organize massive amounts of information to provide to network users. The generation results of large models depend on factors such as pre training data, model parameters, and prompt words input by users. It is difficult for artificial intelligence service providers to predict all the content input by network users in advance. In addition, the amount of generated content is very large, making it impossible to require artificial intelligence service providers to predict, review, and intercept whether the generated content infringes one by one. However, when artificial intelligence generates infringing content due to its own illusions or malicious induction by users, and is notified by the rights holder, the generative AI service provider should and has the conditions to take necessary measures. Therefore, for generative AI service providers, there is a legitimate and reasonable basis for referring to the application of the "safe harbor rules".

For this reason, the "Opinions" have made provisions on the determination of responsibility for the infringement of personality rights by "AI illusions" and the malicious induction of AI infringement by users. It is clear that: if the content automatically generated by generative artificial intelligence infringes on personality rights, and the AI service provider fails to take necessary measures such as stopping the generation of infringing content in a timely manner after being notified by the rights holder, it shall bear tort liability for the damage caused in accordance with the law; If a network user maliciously induces generative artificial intelligence to generate infringing content and causes harm to others by inputting infringing prompt words or other means, the network user shall bear the liability for infringement in accordance with the law. After being notified by the right holder, if the artificial intelligence service provider fails to take necessary measures such as stopping the generation of infringing content and blocking relevant generation instructions in a timely manner, and the right holder requests the network user or artificial intelligence service provider to bear civil liability, the people's court shall support it in accordance with the law. The above regulations comply with the laws of technological development and the legislative spirit of the Civil Code, reflecting the organic unity of supporting innovation and governing by law. During the consultation process, all parties unanimously agreed.

Question: The impact of artificial intelligence is wide-ranging, and judicial activities will also face new risks and challenges. For example, some litigation participants submit evidence and other litigation materials generated by artificial intelligence to the people's court without verification before submission. How does the Opinion respond to the use of artificial intelligence to disrupt judicial order?

Answer: Participants in litigation activities should follow the principle of good faith. With the popularization of generative artificial intelligence technology, it is becoming increasingly common for litigants to use artificial intelligence to assist in litigation. However, artificial intelligence is only an auxiliary tool, and litigants cannot exempt themselves from legal responsibility on the grounds of "AI illusion" and technological neutrality. The parties involved in the litigation shall screen and verify the evidence and other litigation materials submitted to the court to ensure their authenticity and accuracy. In order to guide the public to use artificial intelligence reasonably, prevent false materials from entering the litigation process, and effectively maintain judicial order, the "Opinions" have made provisions from three aspects.

One is to punish the use of artificial intelligence in false litigation in accordance with the law. If a party uses the characteristics of artificial intelligence for autonomous learning and decision-making, and artificially intervenes or misleads by deleting or tampering with synthetic content identifiers, specific instruction inputs, selective result presentation, adversarial interference, etc., to fabricate case facts and engage in false litigation, the people's court shall reject their lawsuit request in accordance with the law, and impose fines or detention according to the severity of the circumstances. If a crime is constituted, criminal responsibility shall be pursued in accordance with the law.

The second is to regulate the use of artificial intelligence to forge evidence in accordance with the law. The Opinion clearly stipulates that if a litigation participant or other person uses artificial intelligence to forge evidence and obstruct the trial of a case by the people's court, they shall be dealt with in accordance with Article 114 of the Civil Procedure Law, that is, they may be fined or detained according to the severity of the circumstances. If a crime is constituted, criminal responsibility shall be pursued in accordance with the law.

The third is to clearly stipulate that litigation participants have the obligation to verify and explain the litigation documents, case retrieval reports, etc. generated by artificial intelligence. At present, the People's Court has discovered multiple false cases generated using artificial intelligence, and the parties and agents have not conducted comprehensive verification before submitting them to the court. In January of this year, news media reported on such incidents in a concentrated manner, and the People's Court case database also included cases. On this basis, the Opinion clearly stipulates that if litigation participants submit litigation documents, case search reports and other materials generated using artificial intelligence, they should carefully verify the authenticity and accuracy of relevant laws, judicial interpretations, cases and other content before submitting them to the court. When submitting them to the court, they should truthfully explain the use of artificial intelligence assistance and bear legal responsibility for the authenticity and accuracy of the relevant content. Regarding this issue, foreign judicial authorities are also highly concerned. We have noticed that some countries have specifically issued guidelines emphasizing the responsibility of agents to ensure the accuracy of materials submitted to the court.



Source: Supreme People's Court