Meta is not alone in grappling with the growing ubiquity of artificial intelligence-generated content and its impact on its platform. In June, YouTube also discreetly implemented a policy change that would enable users to request the removal of artificial intelligence-generated or other synthetic content that mimics their voice or face. This change became effective in June. The modification makes it possible for users to submit a request to YouTube’s privacy request system to have this kind of artificial intelligence material removed. This formalizes the responsible artificial intelligence agenda approach, initially introduced in November and previously articulated.
Rather than requesting the removal of the video due to its deceptive nature, like a deepfake, YouTube encourages the affected parties to directly request the content’s removal, citing it as a privacy breach. According to the most current version of YouTube’s Help literature on the subject, it is necessary to make first-party claims with the exception of a few specific circumstances. These circumstances include those in which the affected person is a juvenile, does not have access to a computer, has passed away, or has other similar circumstances.
However, merely requesting a takedown does not ensure the removal of the content from the website. It is important to note that YouTube will first make its own decision on the complaint based on a number of different considerations. 
For instance, the business may consider whether the information reveals itself as artificial intelligence (AI) or synthetic, whether it enables unique person identification, and whether the content qualifies as parody, satire, or any other valuable and public interest content. Additionally, the business states that it may consider whether the material on the AI platform features a public figure or another well-known human, and whether it portrays the individual in “sensitive behaviour” such as engaging in criminal activities, committing violent acts, or endorsing a product or political candidate. Regarding the latter, this is particularly concerning during election years, as it’s plausible that AI-generated endorsements could potentially sway voting decisions.
YouTube also stipulates that the content’s uploader will have forty-eight hours to respond to the complaint. Before that period of time has elapsed, the complaint will be considered resolved if the material is deleted. On the other hand, YouTube will begin the evaluation process. The firm also warns users that removal involves completely removing the video from the website and, if applicable, removing the individual’s name and any personal information from the video’s title, description, and tags. Moreover, users have the ability to obscure the faces of individuals in their videos; nevertheless, they are unable to simply make the video private in order to comply with the removal request. This is due to the possibility of the video reverting to its public status at any time.
Despite not making any public announcements about the modification of its policy, the firm in March introduced a mechanism within Creator Studio that allowed content creators to report instances of realistic-looking material created through altered or synthetic media, including generative artificial intelligence. In addition, it has only recently started testing a tool that would let users contribute crowdsourced comments that give extra context to videos. These notes could encompass details like whether the video aims to be a parody or if it exhibits any deceptive behavior.
YouTube is not opposed to the use of artificial intelligence, since the platform has previously conducted experiments with generative AI itself. These experiments include a comment summarizer and a conversational tool that allows users to ask questions about a video or get suggestions. However, the firm has previously warned that simply labeling a video as artificial intelligence does not guarantee its removal, as it must still adhere to YouTube’s Community Guidelines.
YouTube is not going to rush to penalize the original content provider in the event that there are privacy issues with artificial intelligence content.
Last month, a representative of the company shared on the YouTube Community website, where the company provides creators with direct updates on new policies and features, that “for creators, if you receive notice of a privacy complaint, keep in mind that privacy violations are separate from Community Guidelines strikes, and receiving a privacy complaint will not automatically result in a strike.”
In other words, YouTube’s Privacy Guidelines differ from its Community Guidelines, and a request for privacy may lead to the removal of some content, even if the video in question does not breach the Community Guidelines. YouTube has informed us that it may take action against accounts that have repeatedly violated the terms of service, even though it will not impose a penalty, such as a limitation on uploads, when deleting a creator’s video due to a privacy complaint.
