Meta responds after Instagram users complain feed is being filled with ‘made with AI’ images

Meta says it is “evaluating” its approach to labelling AI posts after users complained their feeds were being filled with warnings.

As with many other tech companies, Meta’s approach to AI-generated posts on its platforms such as Instagram has largely been to allow it but with a label. The company’s executives have said that is a useful way to allow people to use AI in their work but ensure their audience is aware that it is not necessarily real.

That has led to a warning on Instagram and other platforms that can show along posts, which indicates that the image was “made with AI”. It is automatically generated, though creators can flag their own posts as having used artificial intelligence.

In recent days, however, photographers and others have said that their posts are being flagged as having been made with AI even when the use of such technology is insignificant. Changing even one pixel using an AI-powered tool is enough to trigger the flag.

That led to complaints that users might mistrust a picture that is actually legitimate, simply because an invisible part of it was edited using an AI tool in Photoshop, for instance.

Now Meta says that it is aware of the feedback, and suggested that it could change those requirements in future.

“Our intent has always been to help people know when they see content that has been made with AI. We are taking into account recent feedback and continue to evaluate our approach so that our labels reflect the amount of AI used in an image,” a Meta spokesperson said.

“We rely on industry standard indicators that other companies include in content from their tools, so we’re actively working with these companies to improve the process so our labeling approach matches our intent.”

As with many platforms, Meta relies in part on fingerprinting technologies that allow editing apps such as Adobe’s Photoshop to include information in a file that indicates it has been edited using artificial intelligence. That data is not visible to the person viewing the image – but is included in its metadata, to allow others to understand its provenance.