Meta to Add Broader 'Manipulated Media' Labels, Context to AI Content Starting 2024
-
Updating approach to labeling manipulated media based on feedback from Oversight Board to cover broader range of content beyond just videos.
-
Adding "Made with AI" labels on AI-generated video, audio, images based on detection signals and self-disclosure.
-
Providing more context with labels rather than removing content, unless it otherwise violates policies.
-
Timeline to start labeling AI content in May 2024 and stop removing manipulated media solely based on current policy in July.
-
Policy changes informed by consultations with 120+ global experts, 23,000+ public survey respondents, and input from Oversight Board.