Meta AI

At a glance

Images

Content CredentialsA signed record (C2PA) of how the file was made, sealed so any change shows.

Not stated. None of Meta's pages listed under Sources read says so.

The C2PA statements read concern Meta using Content Credentials to label uploaded AI images on Facebook, Instagram and Threads, not Meta signing Meta AI output.

Invisible watermarkA signal woven into the pixels, frames, audio or wording, meant to survive copying.

Meta's Meta AI media-generation page says Content Seal, an invisible watermark, is on every AI-generated image in Meta AI. On July 7, 2026 Meta said Content Seal keeps a hidden provenance signal in Muse Image images even when they are cropped, compressed, resized or screenshotted.

In February 2024 Meta said photorealistic Meta AI images carried invisible watermarks, without naming them.

Generate images, edit photos, restyle rooms, and more for free with Meta AIMeta (AI at Meta)

Quoted passage FAQ: Is image generation in Meta AI safe?

Yes. Meta AI has built-in safety measures including content policies, prompt filtering, and Content Seal, an invisible watermark on every AI-generated image that provides provenance even after cropping, compression, or screenshots.

Introducing Muse Image and Muse VideoMeta (AI at Meta blog, Meta Superintelligence Labs)

Quoted passage Content Seal

To help people verify whether an image is AI-generated, Muse Image includes Content Seal, our invisible watermarking system. Images created by Muse Image in the Meta AI app and on meta.ai carry a hidden provenance signal that stays intact — even when cropped, compressed, resized, or screenshotted.

Content Seal READMEMeta (facebookresearch on GitHub)

Quoted passage introduction; Content Seal for Images and Video

Content Seal is a comprehensive framework for invisible, robust watermarking across all modalities — audio, image, video, and text." "Content Seal Image is deployed at scale for Muse Image with a custom proprietary implementation. We also provide open-source versions of our research models for images and video, readily available to download.

Labeling AI-Generated Images on Facebook, Instagram and ThreadsMeta (Newsroom)

Quoted passage body, paragraph 2

When photorealistic images are created using our Meta AI feature, we do several things to make sure people know AI is involved, including putting visible markers that you can see on the images, and both invisible watermarks and metadata embedded within image files. Using both invisible watermarking and metadata in this way improves both the robustness of these invisible markers and helps other platforms identify them.

AI label in metadataA plain metadata field (IPTC Digital Source Type) saying AI made the content. Anyone can add or remove it.

In February 2024 Meta said the invisible markers it used for Meta AI images were IPTC metadata and invisible watermarks, in line with Partnership on AI best practices.

Meta did not name the IPTC term, and the 2026 Meta AI pages read do not restate this.

Labeling AI-Generated Images on Facebook, Instagram and ThreadsMeta (Newsroom)

Quoted passage body, paragraph 3

The invisible markers we use for Meta AI images – IPTC metadata and invisible watermarks – are in line with PAI’s best practices .

Visible markA mark you can see, such as a logo or a sparkle in a corner.

In February 2024 Meta said photorealistic images created with Meta AI carried visible markers, and that it applied "Imagined with AI" labels to them.

No Meta page read says whether the label appears on current Muse Image output.

Labeling AI-Generated Images on Facebook, Instagram and ThreadsMeta (Newsroom)

Quoted passage body, paragraph 2

When photorealistic images are created using our Meta AI feature, we do several things to make sure people know AI is involved, including putting visible markers that you can see on the images, and both invisible watermarks and metadata embedded within image files. Using both invisible watermarking and metadata in this way improves both the robustness of these invisible markers and helps other platforms identify them.

Labeling AI-Generated Images on Facebook, Instagram and ThreadsMeta (Newsroom)

Quoted passage body, paragraph 1

We do that by applying “Imagined with AI” labels to photorealistic images created using our Meta AI feature, but we want to be able to do this with content created with other companies’ tools too.

Tool named in metadataThe tool or model written into ordinary file fields, such as Software or Creator.

In February 2024 Meta said photorealistic Meta AI images carried metadata embedded in the image file. It did not list the fields.

The 2026 Meta AI pages read do not restate this.

Labeling AI-Generated Images on Facebook, Instagram and ThreadsMeta (Newsroom)

Quoted passage body, paragraph 2

When photorealistic images are created using our Meta AI feature, we do several things to make sure people know AI is involved, including putting visible markers that you can see on the images, and both invisible watermarks and metadata embedded within image files. Using both invisible watermarking and metadata in this way improves both the robustness of these invisible markers and helps other platforms identify them.

Video

Content CredentialsA signed record (C2PA) of how the file was made, sealed so any change shows.

Not stated. None of Meta's pages listed under Sources read says so.

Invisible watermarkA signal woven into the pixels, frames, audio or wording, meant to survive copying.

Not stated. None of Meta's pages listed under Sources read says so.

On July 7, 2026 Meta said it plans to extend Content Seal to video. It does not say that Meta AI video carries a watermark now.

AI label in metadataA plain metadata field (IPTC Digital Source Type) saying AI made the content. Anyone can add or remove it.

Not stated. None of Meta's pages listed under Sources read says so.

Visible markA mark you can see, such as a logo or a sparkle in a corner.

Not stated. None of Meta's pages listed under Sources read says so.

Tool named in metadataThe tool or model written into ordinary file fields, such as Software or Creator.

Not stated. None of Meta's pages listed under Sources read says so.

Limits and detection

Checking tool from the makerA way the maker offers to check a file for its traces.

Meta's AI Identification page at meta.ai invites users to upload an image, video or audio file to identify files created or edited with AI. On July 7, 2026 Meta said it was previewing a tool to check whether an image carries a Content Seal watermark.

The AI Identification page does not say what it checks for or what its limits are.

AI IdentificationMeta (meta.ai)

Quoted passage page heading

Identify files created or edited with AI" "Upload or drag an image, video, or audio file

Introducing Muse Image and Muse VideoMeta (AI at Meta blog, Meta Superintelligence Labs)

Quoted passage Content Seal

We plan to extend Content Seal to video soon. We’re previewing a detection tool that lets you check whether an image carries a Content Seal watermark, providing an initial way to help you better understand if an image was made with Meta AI.

What the maker says about limitsIn the maker's own words.

In February 2024 Meta said it was not yet possible to identify all AI-generated content, and that people can strip out invisible markers.

Labeling AI-Generated Images on Facebook, Instagram and ThreadsMeta (Newsroom)

Quoted passage body

But it’s not yet possible to identify all AI-generated content, and there are ways that people can strip out invisible markers.