Facebook and Instagram reduce visibility of AI labels in edited content

Meta updates how we label content on Instagram, Facebook, and threads that has been edited or manipulated with generative AI. In an updated blog postMeta has announced that its “AI Info” tag will now appear in the menu at the top right of images and videos edited with AI — instead of directly below a user’s name.

Users can click the dropdown to check if AI information is available and read what may have been customized. Meta previously applied the “AI Information” tag to all AI-related content, whether it was lightly customized in a tool like Photoshop that includes AI features or fully AI-generated from a prompt.

The company says the changes are being made to “better reflect the breadth of AI use” in images and videos on its platforms.

The label was introduced in July after Meta’s previous “Made with AI” label was criticized by creators and photographers for incorrectly tagging real photos they took. “We’ll continue to show the ‘AI info’ label for content we detect was generated by an AI tool and share whether the content was flagged because of signals shared by the industry or because someone self-disclosed,” Meta said in an update, adding that the changes will begin rolling out next week.

The “industry-shared signals” that Meta refers to refer to systems like Adobe’s C2PA-powered Content Credentials metadata, which can be applied to any content created or edited with Firefly’s generative AI tools. There are other similar systems, like SynthID digital watermarks, which Google says are applied to content generated by its own AI tools. Meta didn’t disclose which systems or how many it checks.

Update, September 12: Updated subtitle to clarify that labeling changes for AI-edited content, not AI-generated content.