top of page
  • Writer's pictureSarah Ruivivar

YouTube's New AI-Generated Content Labelling Tool

Updated: Mar 24

Image credits: Szabo Viktor / Unsplash

YouTube has taken a step forward in content transparency by launching a new feature that allows creators to self-label their videos containing AI-generated or synthetic material. This move is part of YouTube's ongoing efforts to maintain authenticity and curb misinformation on its platform.


The new feature, announced today, appears as a checkbox during the video uploading and posting process. Creators are obliged to disclose any "altered or synthetic" content that appears realistic. This includes instances where a real person is manipulated to say or do something they didn't, altering footage of real events or places, or creating a "realistic-looking scene" that didn't actually occur.


However, not all alterations require disclosure. Beauty filters, special effects like background blur, and "clearly unrealistic content" such as animation are exempt from this requirement.


 

Want to learn more about AI's impact on the world in general and property in particular? Join us on our next Webinar! Click here to register

 

This development follows YouTube's announcement in November detailing its AI-generated content policy. The policy established two levels of rules: stricter regulations protecting music labels and artists, and more relaxed guidelines for other content creators.


Under these rules, deepfake music content can be removed by an artist's label if it is deemed inappropriate. YouTube had previously stated that creators would need to disclose AI-generated content but hadn't clarified the process until now.


The new feature operates on an honour system, requiring creators to be truthful about the content in their videos. YouTube has said it is investing in tools to detect AI-generated content, despite the notorious inaccuracy of such software.


In its latest blog post, YouTube stated that it might add an AI disclosure to videos, even if the uploader hasn't, particularly if the altered or synthetic content could potentially mislead viewers. More prominent labels will also appear on videos covering sensitive topics such as health, elections, and finance.


 

Want to learn more about AI's impact on the world in general and property in particular? Join us on our next Webinar! Click here to register

 

Made with TRUST_AI - see the Charter: https://www.modelprop.co.uk/trust-ai

31 views0 comments

Recent Posts

See All

Comments


bottom of page