YouTube is taking a more active role in policing AI-generated content, announcing it will automatically apply labels to videos when its internal systems detect significant photorealistic AI — even when creators fail to disclose it themselves.
The platform has required creator disclosure for over two years, but the policy shift means YouTube will now act unilaterally if that obligation goes unmet. Videos made with YouTube’s own AI tools, including Veo and Dream Screen, will receive permanent labels that creators cannot remove. Content carrying C2PA metadata confirming full AI generation will also be labelled automatically.
The move follows Google’s unveiling of Gemini Omni at Google I/O, a multimodal model capable of producing high-quality AI video. Labels will also become more visually prominent, appearing directly below the video player on long-form content and overlaid on YouTube Shorts.
