YouTube is setting new requirements for creators to label videos produced with artificial intelligence that look realistically generated, starting this Monday. This move aims to maintain transparency and prevent viewer confusion or deception by synthetic content.
Creators will encounter a checklist during the upload process to indicate whether their video portrays a real individual performing actions they did not do, modifies footage of actual places or events, or showcases a scene that appears real but never occurred.
This initiative addresses the growing concern among online safety experts over the potential for AI-generated content to mislead internet users, particularly with the upcoming 2024 elections in the United States and other countries.
Failure to identify videos with AI-generated or manipulated content that seems realistic will prompt YouTube to add a viewer notice. This disclosure will signal “altered or synthetic content” within the video description, emphasizing significant edits or digital generation of sound and visuals. For videos touching on "sensitive" issues like politics, this label will appear more conspicuously on the video screen.
The policy also extends to content created using YouTube's own AI tools, introduced last September, ensuring it is explicitly marked.
However, YouTube will not require labels for synthetic or AI-generated content that is clearly fictitious or "inconsequential," such as AI-powered animations or minor visual adjustments. The platform also exempts the disclosure of AI's role in enhancing productivity through script generation, content brainstorming, or automated captions.
Creators neglecting to apply the necessary label on relevant synthetic content risk sanctions, including video removal or suspension from YouTube's Partner Program, which allows for content monetization. This regulation is part of YouTube's upcoming suite of AI policies scheduled for a fall release.
