What is AI info on Facebook and Instagram? Meta's AI label

"AI info" is the label Meta puts on Facebook, Instagram and Threads posts whose video, audio or images were made or edited with AI. Meta adds it when it detects industry-standard AI indicators in the content, or when the person posting discloses that it is AI-generated, and people can click the label for more information.

Meta's rules here are quoted from two Meta Newsroom posts, Labeling AI-Generated Images (February 2024) and Our Approach to Labeling AI-Generated Content (April 2024, with later updates), read on 2026-09-28. Meta's Help Center pages don't render without JavaScript, so they aren't quoted. Sume facts come from its Image API docs and Terms.

Why does Facebook label a picture as AI?

Either someone disclosed AI when posting, or Meta detected an indicator in the file. Meta says it detects the "AI generated" information in the C2PA and IPTC technical standards, so it can label images from other companies' tools as they add that metadata. Images made with Meta AI get visible markers plus invisible watermarks and metadata.

The label used to say "Made with AI". Meta found that some content with minor AI modifications, such as retouching tools, carried those indicators and was labeled "Made with AI", so in July 2024 it renamed the label "AI info". In September 2024 it said it would move the label into the post's menu for content it detects was only modified or edited by AI, and keep it visible for content it detects was generated by an AI tool.

Is it necessary to add an AI label on Instagram or Facebook?

For realistic video and audio, yes. Meta says it will require people to use its disclosure and label tool when they post organic content with a photorealistic video or realistic-sounding audio that was digitally created or altered, and that it may apply penalties if they fail to do so. It gave the reason: other companies' tools weren't yet adding detectable signals to audio and video at the same scale as images.

Will Meta remove a post with an AI info label?

Not for the label alone. Meta says it keeps labeled content on its platforms unless it otherwise violates its policies. Meta's April 2024 post adds that independent fact-checkers review false and misleading AI-generated content, and that a post they rate False or Altered is shown lower in Feed with an overlay label. For ads, Meta says advertisers have to disclose when they digitally create or alter a political or social issue ad in certain cases; its other ad rules are outside the pages quoted here.

Will Meta label a video or image made with Sume?

Sume's docs don't say whether its files carry C2PA or IPTC markers, so this post can't predict an automatic label. If a Sume video looks real, disclose it yourself when you post. Sume outputs this can apply to include images from POST /v1/images (Image API) and avatar talking videos, which turn a ready avatar into a script-driven talking video.

Sume's Terms point the same way: generated outputs are synthetic and should not be treated as a factual record of real people or events, you need permission to use any person's likeness or voice you submit, and you must not use Sume to mislead people about the authenticity of generated media. For file specs, see Meta video ad specs.