YouTube Introduces AI-Generated Content Labeling Tool

What to know

• YouTube now requires creators to disclose when their videos contain altered, synthetic, or AI-generated content.
• Creators can label videos that have ‘Altered content’ from YouTube Studio’s ‘Details’ page.  
• Videos with altered but meaningfully realistic content that have the potential to mislead or confuse people will have to be labeled as ‘altered’. 
• Videos with inconsequential usage of synthetic media, such as effects and filters, or those that are obviously not real, such as animation, don’t require the label.

Keeping up with greater transparency trends, YouTube has added a new tool in YouTube Studio that will allow creators to disclose when videos include altered or synthetic media, as well as AI-generated content.

Creators can label their videos as containing ‘Altered content’ from the Details section while uploading their videos.

Creators are encouraged to disclose content that:

• Makes a real person appear to say or do something they didn’t do.
• Alters footage of a real event or place.
• Generates a realistic-looking scene that didn’t actually occur. 

This includes any content that is fully or partially altered using image, audio, or video editing tools, as well as content that is generated by AI.

Once a creator marks their video as having altered or synthetic content, the label will be shown as part of the video descriptions. 

This is to encourage honesty and trust between the creators and the audience while reducing deep fakes and misinformation.

However, do note that YouTube isn’t forcing creators to label their videos as such. This is because creators use generative AI in several ways, and not all of them for nefarious purposes. Videos where synthetic media use is inconsequential, such as when using filters, effects, or visual enhancements, or videos that are clearly unrealistic such as animations are not required to label their content as such. 

On the other hand, YouTube has made it clear that any video that uses altered content that could be misconstrued as the real thing ought to have the label. The only reason why it isn’t mandatory yet seems to be because of the difficulty and unreliability of AI detection tools.

YouTube’s blog does go on to state that YouTube will “look at enforcement measures for creators who consistently choose not to disclose this information. In some cases, YouTube may add a label even when a creator hasn’t disclosed it, especially if the altered or synthetic content has the potential to confuse or mislead people.” 

So there are real risks of not disclosing altered content, especially if it is meaningfully realistic, and could very well lead to “penalties from YouTube, including removal of content or suspension from the YouTube Partner Program”, as mentioned on Google’s support page.