Meta is updating the way AI-curated content is tagged across Instagram, Facebook, and Threads. AI-curated images and videos will now have the “AI Insight” label appear in a menu in the top right corner of content, rather than just below the user name. The change was announced in a blog post by Meta.
By clicking on this menu, users will be able to learn whether the content has been edited with AI and what kind of changes have been made. Previously, Meta added an “AI Knowledge” label to content that had been lightly edited with AI-enabled tools like Photoshop or content that was entirely produced with AI. The company stated that this change was made to more accurately reflect the extent of AI used in images and videos.
Why is Meta updating its AI tags?
The social media giant has decided to update this labeling system, which it introduced in July, after criticism of its previous “Created with AI” label. Photographers and content creators had complained that real photos they took were incorrectly labeled as AI-generated. “We will still add the ‘AI Informed’ label to content we detect as created with an AI tool, and will indicate whether the content was tagged using industry-shared signals or self-reports,” the company said in its update. The company added that these changes will start rolling out next week.
Among the “industry-shared signals” Meta mentioned are C2PA-supported Content Credentials metadata that Adobe applies to content created with its Firefly AI tools. Similarly, Google has announced that it is adding digital watermarks called SynthID to content created with its own AI tools. However, Meta did not provide specific information about which systems it controls and how much of them.
On the other hand, completely removing labels from real AI-edited images could make it harder to protect users from misleading content. It could make it harder for users to spot manipulated images, especially as AI editing tools on new phones become increasingly convincing.