Why did Instagram add an AI info label to my photo?
Meta labels images on Instagram, Facebook and Threads when people disclose AI use or when it detects industry-standard AI indicators. It has said it reads the “AI generated” information in the C2PA and IPTC standards, which tools from Adobe, Google, OpenAI, Microsoft, Midjourney and Shutterstock write into the files they produce. An editing app can write those records even when most of the picture came from your camera: Adobe says Photoshop’s Generative Fill supports Content Credentials that show whether content was AI-generated or AI-edited.
The records are part of the file, so they travel with every export of it. A label can also come from a disclosure or from detection that does not depend on the file, and the checker cannot see either of those.
Source: Meta: labeling AI-generated images (opens in a new tab)
Check the file before you post again
Run the check on the same file you upload, not on a screenshot or on a copy saved back from Instagram, which has usually lost these records.
- Export the photo the way you normally would, then choose that file in the checker above. It is read in your browser tab and is not uploaded.
- Content Credentials found: the evidence details give the organization that claims to have signed them, such as Adobe, and say whether they declare AI generation or an AI edit.
- AI generation declared: the file’s metadata or credentials name an AI source type. In the IPTC vocabulary compositeWithTrainedAlgorithmicMedia covers an AI edit such as inpainting and trainedAlgorithmicMedia an image created by an AI model; the checker reports either one as declared AI.
- Nothing found: this file carries no AI record the checker can read, so the label probably came from a disclosure setting or from Meta’s own detection.
- Compare an export made before the AI edit with one made after it to see which step added the record.
Source: IPTC Digital Source Type vocabulary (opens in a new tab)
What does the AI info label mean now?
On 1 July 2024 Meta renamed its “Made with AI” label to “AI info” after minor AI edits such as retouching triggered the old label. Since 12 September 2024, content Meta detects as only edited or modified with AI tools shows the label in the post’s menu, while content it detects as generated by an AI tool keeps the label on the post. Tapping the label shows whether it came from industry signals or from the poster’s own disclosure.
If your file does record an AI edit
Then the label reflects what your editing app declared. Removing those records to avoid a label also removes the evidence that only part of the picture was changed, and Image Evidence does not offer metadata removal or detector evasion. Saying in the caption what you edited, and how, keeps the context with the post. For edits that leave no record, our guide to AI-edited photos covers other checks.
What the checker cannot tell you
It reads one image file that you choose. It does not open Instagram posts, stories or profiles, and it cannot see Meta’s detection results, invisible watermarks such as SynthID or Meta’s Content Seal, or the disclosure setting on a post. A clean result shows that the file carries no AI record the checker can read; it does not prove that a photo is unedited. For checking someone else’s post, see Is this Instagram post AI?.