{"id":6515,"date":"2026-09-04T12:06:47","date_gmt":"2026-09-04T12:06:47","guid":{"rendered":"https:\/\/silvybrand.com\/?p=6515"},"modified":"2026-09-04T12:06:47","modified_gmt":"2026-09-04T12:06:47","slug":"instagram-ai-content-label-confusion","status":"publish","type":"post","link":"https:\/\/silvybrand.com\/?p=6515","title":{"rendered":"Instagram\u2019s AI detection is a mess (again)"},"content":{"rendered":"


\n<\/p>\n

\n
\n

Instagram\u2019s visible AI labels are supposed to help people quickly spot synthetically generated content at a glance. Over the last few weeks, however, users<\/a> have been reporting that the system has gone haywire. They say Meta has been automatically applying an \u201cAI Content\u201d label to images that they didn\u2019t create or edit using generative AI tools. Meanwhile, actual AI imagery is slipping through the cracks, leaving the impression that nothing can be trusted on Instagram at all.<\/p>\n<\/div>\n

\n

The precise causes of this tagging seem to vary. Many people said the label appeared on images that were edited using tools like Canva\u2019s Background Remover or with negligible use of blemish-fixing tools. One user on Threads<\/a> said Instagram was applying the label \u201cevery time there is bg remover involved,\u201d and another user<\/a> said it appeared after they used Canva to remove \u201ca speckle\u201d from their photo, resulting in the entire image being flagged on Instagram as AI. In other cases, the causes are even less clear.<\/p>\n<\/div>\n

\n

If this labeling mishap sounds familiar then you may recall that a similar issue blighted photography on Instagram with a \u201cMade by AI\u201d<\/a> label in 2024, a few months after the feature was released. That February, Meta announced it would scan images for IPTC and C2PA metadata<\/a>, which can prove with reasonable certainty whether generative AI was used to create or manipulate them. But it\u2019s always been vague about its methods. In 2024, the detection system reportedly swept up pictures whose Adobe metadata indicated they had been retouched with generative AI tools, even if the changes were extremely minor and substantively resulted in the same photograph. Meta then promised to tweak its labeling approach to better reflect \u201cthe amount of AI used in an image.\u201d It also said it uses \u201cindustry standard indicators that other companies include in content from their tools,\u201d but there isn\u2019t any recent information available about what they are, or how and when it scans for them. Meta didn\u2019t respond to a request for clarification.<\/p>\n<\/div>\n

\n
\n
\n
\n
\n
\"A<\/a><\/div>\n<\/div>\n<\/div>\n

This is the message that pops up when you inspect the AI Content tags that Instagram is automatically applying.<\/em><\/figcaption>Image: Meta<\/cite><\/p>\n<\/div>\n<\/div>\n<\/div>\n
\n

The recently reported erroneous labels don\u2019t have a clear link to generative AI usage. While features like background removal do<\/em> often use machine learning, it\u2019s the same assistive AI variety that apps like Photoshop have used in object selection and removal tools for over a decade \u2014 not the generative text-to-picture tools typically associated with deceptive AI imagery.<\/p>\n<\/div>\n

\n

At least some of Instagram\u2019s recent AI labeling mishaps appear to be Canva\u2019s fault. After flagging the Background Remover tagging concerns, content strategist Jess Bruno<\/a> reports that Canva came back to her with an explanation: Some of the design platform\u2019s assistive AI tools \u201cwere being tagged as generative.\u201d The design platform told Bruno that its tools are now tagging correctly. A message on Canva\u2019s background removal help page<\/a> also says that using the tool \u201cdoesn\u2019t add Canva\u2019s AI-generated content metadata to your design,\u201d though it\u2019s unclear if that note is new, and Canva didn\u2019t respond to a request for comment.<\/p>\n<\/div>\n

\n

Yet some Threads users<\/a> are reporting that using Canva\u2019s Background Remover still results in their images getting tagged, despite Canva claiming to have resolved the issue on its end. Other Threads<\/a> users also report that none of the images they edited using the tool prior to it being fixed got tagged by Instagram, nor did the tool apply C2PA metadata that Meta supposedly uses to detect AI-generated content.<\/p>\n<\/div>\n

\n

Out of all the tests I ran to see what would make the label appear on Instagram, it was only applied on images I edited or fully generated using the Meta AI app.<\/p>\n<\/div>\n

\n
\n
\n
\n
\n
\"12<\/a><\/div>\n<\/div>\n<\/div>\n

These are just some of the images I posted to Instagram to test its AI detection, but only the two on the top right (created in Meta AI) were tagged.<\/em><\/figcaption>Image: Jess Weatherbed \/ The Verge<\/em><\/cite><\/p>\n<\/div>\n<\/div>\n<\/div>\n
\n

There are also reported cases of AI tagging<\/a> where Canva\u2019s Background Remover was reportedly never used at all<\/a>. One user found that Meta applied the tag on an image that had been poisoned<\/a> \u2014 or subtly tweaked by a system that\u2019s designed to make images useless for AI training and\/or degrade the AI models they\u2019re fed to \u2014 and didn\u2019t tag an unpoisoned version of the same content.<\/p>\n<\/div>\n

\n

The AI Content tag was also automatically applied to recent photos posted to Instagram by About Face, a cosmetics company founded by the singer Halsey. When a commenter inquired about the AI tag<\/a>, the brand\u2019s unnamed social manager said that no AI was used to create the images. \u201cThis photo was taken on my iPhone and then slightly edited by myself in the photo app. Our brand uses real artist[s] and real people to create everything you see across all our platforms,\u201d reads the About Face reply.<\/p>\n<\/div>\n

\n

I reached out to About Face to clarify which tools in the iPhone Photos app were used, but the company has yet to respond. That said, only a handful should add generative AI metadata to photographs: Spatial Reframing, Extend, and updated Clean Up added in iOS 27<\/a>. These tools use Apple Intelligence and get embedded with Google\u2019s invisible SynthID watermark; the first two add new generative elements, while the third lets users remove distracting objects or people. The tagged images don\u2019t have SynthID embedded, according to Google\u2019s Gemini-based verification, but even if those Apple Intelligence tools were used, it\u2019s not clear why the label says it detected signals that \u201ctypically identifies content that was entirely AI-generated instead of content that was only modified with AI.\u201d<\/p>\n<\/div>\n

\n

Seriously, what AI signals is<\/em> Meta actually scanning for? Companies often obscure exact details of their moderation systems to prevent bad actors from gaming the system, but right now, ordinary users are reasonably confused.<\/p>\n<\/div>\n

\n
\n
\n
\n
\n
\n
\n
\"\"<\/a><\/div>\n<\/div>\n<\/div>\n

These two images are entirely generated in Google Gemini using reference photographs of me (the likeness of which is questionable).<\/em><\/figcaption>Image by Jess Weatherbed \/ The Verge, generated in Gemini<\/cite><\/p>\n<\/div>\n<\/div>\n
\n
\n
\n
\n
\"\"<\/a><\/div>\n<\/div>\n<\/div>\n

I have never visited a snow-covered lodge, worn this outfit, or had my hands tattooed. Both images have C2PA and SynthID embedded, and Instagram hasn\u2019t tagged either of them as AI.<\/em><\/figcaption>Image by Jess Weatherbed \/ The Verge, generated in Gemini<\/cite><\/p>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n
\n

The only signal that I definitely know will currently trigger Instagram to apply an AI Content label is the one applied by Meta\u2019s own generative AI systems. I ran my own tests of most of the tools above, uploading photos that were edited using Canva\u2019s Background Remover, Photoshop\u2019s background erasing tool, Adobe\u2019s Firefly generative AI features, Google\u2019s Nano Banana model in Gemini, and iOS Apple intelligence features. Some were fully generated, and I confirmed that several of these included C2PA and SynthID signals. In my case, however, Meta didn\u2019t label any of it.<\/p>\n<\/div>\n

\n

I posted the pictures on a brand-new Instagram account I set up with minimal information, and I published them in quick succession. I couldn\u2019t have tried harder to act like a bot or AI farm, which you\u2019d think would be a prime contender for what Meta should be looking for. Yet even after Meta announced crackdowns on AI accounts last week, it wasn\u2019t tagged as an AI-generated profile, nor were the pictures labeled<\/a> then. At this point, they\u2019ve been live and fully public for almost two weeks.<\/p>\n<\/div>\n

\n

The results have left me questioning if there is<\/em> any sense in how Meta is applying \u201cAI Content\u201d tags. They\u2019re supposed to make it easier to trust what you see on Instagram, but at this point, I wouldn\u2019t even trust the labels themselves.<\/p>\n<\/div>\n

Follow topics and authors<\/strong> from this story to see more like this in your personalized homepage feed and to receive email updates.<\/span><\/p>\n