YouTube to begin automatically labeling AI videos

YouTube has announced it will begin automatically applying labels to some AI videos, with the key point being that when the platform detects "significant, realistic" AI-generated or AI-modified content, YouTube will apply the label on the creator's behalf even if the creator did not proactively disclose it. This means YouTube's AI content disclosure mechanism is shifting from being more reliant on creators filling in the information themselves toward the platform actively detecting and supplementing labels. For viewers, the core value of this change is reducing the risk of mistaking synthetic imagery for real people, places, or events—especially as AI video models become increasingly realistic, where being able to see a clear label at the moment of viewing affects judgments of trust. YouTube will also make AI labels more visible: on long-form videos the label will appear below the player and above the description box; on Shorts it will be presented as an overlay. This is more direct than hiding it only in the expanded description and is closer to a design of "real-time prompting while consuming content." However, this update does not mean all AI content will be prominently labeled. The report specifically notes that animations, content that is clearly not realistic, or videos with only minor AI involvement may still not clearly reveal their AI origin, or may disclose it only in a less prominent location. In other words, YouTube's new approach mainly targets realistic AI content that could be mistaken for real, rather than uniformly labeling all AI-assisted creations the same way. For creators, independent media, and brands, this raises transparency requirements for AI visual content, but also leaves gray areas: what counts as "significant" or "realistic" AI will still depend on how the platform's enforcement and appeals mechanisms are implemented going forward.