The Beijing Internet Court (BIC) recently heard a case concerning reputation infringement arising from the publication of AI-generated content.

Case summary

The plaintiff, an agency engaged in live-streaming services, attracted public attention after one of its streamers unfortunately passed away. The defendant posted a video on their social media account, claiming that the streamer "relied on medication and alcohol to get by, suffering from insomnia and anxiety," "live-streamed 15 hours a day and stayed up late memorizing scripts," "carried antidepressants and passed them off as throat lozenges," and that "doctors advised rest and treatment, but the team pushed for more live-streams."

The plaintiff argued that these statements were fabricated and damaged its reputation, and filed a lawsuit requesting that the defendant cease the infringement, issue a public apology, and compensate it 30,000 yuan ($4,434) for economic losses and reasonable expenses.

During the proceedings, the defendant deleted the video and argued that the script was not original but generated by AI. The defendant submitted screen recordings showing that an AI tool had been instructed to "write a script about the streamer's death, advocating health awareness," and was later asked to supplement the script with online reports.

After retrieving multiple articles from several platforms and revising the script, the AI generated the final version, which the defendant used directly to produce and publish the video. The defendant contended that, as the content was derived from publicly available information, it did not constitute reputation infringement

After trial, the BIC determined that:

The "streamer" mentioned in the video was under contract with the plaintiff, and the "team" referred to in the video was the plaintiff agency. Although the defendant argued that the content was AI-generated and based on publicly available reports, the BIC held that the defendant had a duty to verify the accuracy of the information before using AI-generated content to produce and publish the video.

The defendant neither identified the sources of the information or verified its authenticity, and failed to provide evidence supporting statements such as "live-streaming 15 hours a day," "staying up late to memorize scripts," or "doctors advised rest but the team pushed for live-streams."

After the video's release, numerous social media users made negative comments about the plaintiff, damaging its social reputation. The court therefore held that the defendant had infringed upon the plaintiff's reputation and should bear corresponding liability.

Details of the judgment

The BIC ordered the defendant to publish an apology through the social media account used to post the involved video and to compensate the plaintiff for certain economic losses. The judgment has now taken effect.

Comments by the judge

Generative AI has become an important tool for obtaining information, supporting cognition, creating content, and supporting decision-making. At the same time, AI-generated content has inherent limitations and may factual inaccuracies, logical fallacies, and even AI "hallucinations."

If used or disseminated without verification, such content can mislead the public, cause erroneous judgments and decisions, and lead to disputes and risks. Content users, publishers, and relevant platform operators all bear primary responsibility and a statutory duty to verify the authenticity of information. They cannot evade liabilty on the grounds that the content was generated by AI. Before publishing or reposting any text, image, audio, or video content, the facts and sources must be verified to avoid infringing upon others' reputation, privacy, copyright, and other lawful rights and interests.