YouTube recently announced in a blog post that in the next couple of months, it will allow people to submit removal requests for AI-generated or other synthetic or altered content that simulates an identifiable individual, including their face or voice (deepfakes). The company states that all reported content will not be taken down, and that factors, such as whether the content was parody or satire, will be considered by the company when considering removal requests. It will also consider whether the person making the request can be uniquely identified or a well-known individual in which case, “there may be a higher bar.” The company does not clarify what this higher bar will be, but this could be interpreted to mean that the evaluation process for removal requests will be more rigorous in cases involving deepfakes of well-known individuals.
YouTube also said that its music partners will be able to submit removal requests for AI-generated music that mimics an artist’s unique singing or rapping voice. For these removal requests the company will consider factors such as whether content is the subject of news reporting, analysis, or critique of the synthetic vocals.
Labeling of AI-generated content
Another feature that the company mentioned was that it will put in place content labels that inform viewers if a piece of content has been AI-generated. The onus of labeling content as synthetic or AI-generated will be on the creators. Notably, the creators will be required to label only those pieces of altered or synthetic content that look realistic. In AI-generated videos that discuss sensitive topics such as elections and ongoing conflicts, these content labels will be prominently displayed in the video player. For less sensitive content, the label will be placed in the video description.
If a content creator consistently fails to label their AI-generated content, they may be subject to penalties such as content removal and suspension from the YouTube Partner Program. YouTube says that in cases where a piece of AI-generated content violates the company’s Community Guidelines, it will be removed from the platform.
How the content labels would be displayed
Why it matters:
It is important to point out here that content creators will have to use their discretion in deciding which piece of synthetic/ AI-generated content will be labeled. This means that if the creator deems that a piece of AI-generated content isn’t “realistic” (but viewers think looks real), it will freely circulate on the platform. While the company says that it will hold creators accountable by content labeling with measures like content removal, the company has not clarified how it will identify AI-generated content. The lack of clarity on YouTube’s AI identification measures leaves one wondering whether these measures will be enough to tackle the threat of misinformation posed by AI content.
- Microsoft To Help Combat Deepfakes In The Run Up To 2024 Elections
- Adobe Is Selling AI-Generated Images Of The Ongoing Conflict In Gaza And Israel
- No Generative AI Tools For Political Advertisers: Meta
STAY ON TOP OF TECH NEWS: Our daily newsletter with the top story of the day from MediaNama, delivered to your inbox before 9 AM. Click here to sign up today!