Some users complain that Meta is tagging real photos with its “Made with AI label”, possibly because AI tools were used to edit, rather than create, the images
TechCrunchIvan Mehta
Context & Ripple Effects
Meta’s complaint-driven labeling problem exposes a boundary its earlier plan for industry standards to identify AI content needed to resolve: whether a label describes an image’s origin, its editing history, or both.
The distinction mattered quickly enough that Meta later renamed the label “AI info” to signal that AI involvement did not necessarily mean an image was wholly generated. That shift makes the initial wording a consequential product-design and trust issue, not merely a cosmetic label change.
First-order effects
Photographers and other users whose real images were edited with AI tools can be publicly associated with AI-generated content even when the image was not created that way.
Meta’s “Made with AI” label becomes less precise for viewers, while the company must distinguish creation signals from edit-related signals in its disclosure experience.
Second-order effects
Creators may need to explain or contest labels on authentic work, adding friction where audiences or clients treat AI generation and AI-assisted editing differently.
Platforms pursuing shared AI-content identifiers face pressure to make provenance disclosures more granular; a single binary label can conflate distinct forms of AI involvement.
Third-order effects
If editing tools become routine across image workflows, provenance systems will need to communicate a spectrum—from capture through modification—rather than imply a simple real-versus-synthetic divide.
The episode points toward a synthetic-media control plane in which labels are judged not only on whether they detect AI signals, but on whether users can understand the specific claim being made.
The trend: AI-content disclosure is shifting from broad generated-content labels toward more nuanced provenance signals that separate creation from editing.
The blanket requirement to disclose media made with AI is reactionary, unnecessary, and destined to lead to situations like this; because where do we draw the line at how much use of AI (which is what Generative Fill is, lest we forget) is the limit for ‘real’ or ‘fake’? 1%? 10…
I can't read mentions of AI labeling without wondering why nobody in the media is talking about how it's going so far two months in. 1. Not one single photographer nor artist on all of Facebook and Instagram has the first clue what triggers the “Made with AI” label. Even though…
My first brush with the dreaded “made with AI” tag. Not a single bit of generative fill used anywhere on any of these (because I f**king hate AI) so that's a tad concerning 🤔
@mosseri the inability for @instagram to distinguish between using ai to generate photo realistic images or using generative ai to aide in time saving or complex removals does a disservice to photographers AND viewers.
Instagram automatically tagging a photo “Made with AI” when it detects generative fill is wrong and bad and will ultimately do more harm than good. If “retouched” photos are “Made with AI” then that term effectively has no meaning. They might as well auto tag every photograph “…
This is outrageous! — Also, heads up #kidlit that Meta may now (on top of everything else 🙄) label your art as A.I. if you create it with an Adobe product. — My vote for word of the year? Sloppy. [embedded post]
Love Facebook's “made with AI” label — Watching it force a label on *legitimate photographs* but skip all of the terrible groups their algo keeps suggesting to me with OBVIOUS clickbait AI images is just... Why even pretend you're trying?
These Meta false positives for “Made with AI” are so messy and detrimental to creatives whose work is genuine (or edited via Photoshop etc and not AI generated) Just another product-based example of the Tech/Creative divide. Get it right before you ship the feature.
Oh, nice— now all my posts are not only labeled as “Made with AI,” but they also are simply invisible. Refreshing does nothing. Good times, and thanks for being terrible, Meta. 🤬 [video]
@kyletwebster Happened with us when we released our summer campaign, which used in-house photography and graphic design. Put a dent in our optics and brand image, not to mention in our soul too.
Apparently Instagram's Meta just DECIDES when something is AI and labels your post as such with no option to remove it??? And no care for how it could harm an artist's reputation to label their art as AI when it is not. I'VE NEVER USED AI AND NEVER WILL. GET FUCKED INSTAGRAM [ima…
I think it's way more likely that their AI detection tool is just not that accurate. I don't think it has anything to do with what people do. It's just bad AI falsely detecting things it shouldn't.
Instagram has assigned the label “Made with AI” on ALL content across the board when they're not made with AI, rendering all users' pages with blank squares. @instagram @Meta [image]