Meta announced earlier this week that it would be changing its language regarding artificial intelligence after complaints from photographers who had noticed their photos were labeled “created with AI.”
Meta introduced the “Made with AI” label earlier this year following a review of Meta's policies by the Federal Communications Commission, which concluded that artificial intelligence images were indistinguishable from real images.
Images mislabeled as AI-generated
Photographers found that real photos were being falsely labeled as AI-generated. An investigation conducted by Meta found that images that had been retouched or otherwise manipulated using various AI tools were being marked as such. Meta's computer systems recognized that real images had been manipulated and affixed AI labels to them, leading to complaints from photographers.
Former White House photographer Pete Souza was one of the complainants, saying one of his photos had been labeled by artificial intelligence. He expressed suspicion that processing the image with Adobe Photoshop software may have led Meta's algorithm to add the tag.
This week, Meta announced that it has revised the labels it adds to images that it identifies as being created with artificial intelligence. By clicking on the “AI Info” label, users can view more information. When this happens, text will appear stating that Meta may add a label if content is shared that the system recognizes as containing artificial intelligence.
