YouTube will apply visible 'AI' labels to photorealistic or meaningfully altered content starting in May, moving disclosures from video descriptions to placements below long-form players and overlays on Shorts. The changes build on existing policies in place for over two years and coincide with Google's release of Gemini Omni models. Both TechCrunch and The Verge report the updates rely on automated internal signals rather than creator self-disclosure alone.