YouTube Will Now Automatically Label AI-Generated Content, Deepfake Protection for All 18+ Creators Also Announced

Google’s YouTube has now announced new updates for AI-related contents on the platform, and with this, more visible AI labels will be brought out for viewers. Along with this, a new internal checking system will also be introduced to automatically label certain AI contents, while the likeness detection technology will be rolled out for creators in the coming weeks.

Read more about these latest updates below.

YouTube’s New Visible AI Labels and Deepfake Protection

First among the new updates is the more visible AI labels for videos. Earlier, creators who used AI tools for creating or meaningfully altering their videos had to disclose it, and the label was shown inside the description section. Now, as per the details shared, YouTube will be moving this label to a more prominent position, thanks to which viewers will be able to easily know when they are watching photorealistic or meaningfully AI altered/generated content.

For long-form videos, the label will now appear directly below the video player and above the description. For Shorts, the label will be shown as an overlay on the video itself. To note, if the content is clearly animated, unrealistic, or only slightly edited, the AI disclosure will still be placed inside the expanded description section.

Via the automatic AI detection system, YouTube will be able to add AI labels on its own when significant photorealistic AI content is detected. However, creators are still required to manually disclose realistic AI content while uploading their videos, but if they forget or do not add it, YouTube’s internal system will automatically apply the label when needed.

In case the video has been wrongly labelled, YouTube will still allow creators the option to fix or remove the label from YouTube Studio. However, the AI label cannot be removed if the video has been made using YouTube’s own AI tools, such as Veo or Dream Screen, or if the video contains C2PA metadata that proves it to be fully generative AI content.

The likeness detection technology is another key update, and this will help creators to spot AI deepfakes of themselves. This feature will be made available to every channel owner who is 18 or older, and it won’t matter whether the channel has subscribers, monetisation enabled, or even any uploaded videos. Eligible creators will be getting access to it over the next few weeks.

Furthermore, as clarified by YouTube, having an AI disclosure label on a video will not affect how the video is recommended on the platform, nor will it affect the monetisation eligibility of the video. Google adds that this label is only being brought out for transparency, so that viewers can know when AI tools have been used in a video.

Stay tuned for more updates.