Deceptive AI or edited media of real people or events
X may show the video to fewer people.
Last verified
What triggers it
Media shared as real that is significantly and deceptively altered or fabricated: AI-simulated real people, new frames, overdubbed audio or changed subtitles that change the meaning, or edits that distort what happened, where it could cause widespread confusion on public issues, affect public safety or cause serious harm.
What X says
“media depicting a real person that has been fabricated or simulated, especially through use of algorithms or broader artificial intelligence.”
We last checked this against X's page on .
How to fix it
Disclose AI or edited parts clearly in the caption and on screen (and turn on the Made with AI label), or remove the fabricated parts.
Good to know
Watch for realistic AI faces or voices of real public figures saying or doing things, and cloned voices. Allowed when not deceptive: clearly labeled AI, memes, satire, animation and commentary that do not cause significant confusion about authenticity. Retouching and color correction do not count.
Can FYProof spot it?
A full check of your video looks for it in what's shown, what's said and the caption. What a check finds is an estimate, not a guarantee.
- Video: looked at for this rule
- Speech: looked at for this rule
- On-screen text: not for this rule
- Caption: looked at for this rule
- Hashtags: not for this rule
Sounds other than speech, such as music, aren't checked.
It doesn't read details stored inside the file, such as labels an editing app adds.
Related X rules
Platforms change their rules often. This page sums up what X publishes; it isn't legal advice.