YouTube mandates AI-generated video labels

YouTube has told content creators to start labeling videos that have been created with artificial intelligence (AI) and that look realistic.

YouTube said in a statement: “Generative AI is changing the way creators tell stories, but viewers want to know when the content they’re watching has been created with AI.”

YouTube is trying to require creators to inform viewers about content that has been manipulated or created with AI tools that looks real and could easily be mistaken for a real person, place, or event.

YouTube gave the following examples of what should be flagged as a manipulated video:

  • Resemblance to a real person: Digitally altering content to replace one person’s face with another, or artificially generating a person’s voice to narrate a video.
  • Video of real events or places: Making it appear as if a real building is on fire, or changing a real cityscape to look different than it actually does.
  • Creating realistic scenes: Showing a realistic-looking depiction of fake events, such as a tornado heading towards a real city.

If the content creator flags the video, a special icon will be displayed to the viewer indicating its authenticity. YouTube says that for most videos, this label will appear in the video description, but for content that covers more sensitive topics – such as health, news, elections or finance – the label will appear on the video itself.

YouTube added that it will not require creators to flag content that is clearly non-realistic or has an animated style.

YouTube does not require creators to disclose their use of AI for video production; for example, there is no need to mention AI when it is used to create video scripts, content ideas, or automatic subtitles.

The new labels will start appearing on all versions of YouTube in the coming weeks.


Back to top button