Meta’s AI Photo Label Tagging Real Photos by Mistake
Editing Isn't AI!
Meta, the company behind Facebook and Instagram, has pledged to classify images generated using artificial intelligence (AI) features. Since May, they’ve been wrongly labelling real images with “Made with AI” labels.
Users and photographers have been angered as their unedited photographs have been labelled as AI. Pete Souza, a former White House photographer, even had one of his images flagged! He believes it is due to a recent modification in a popular editing tool (Adobe), which triggers Meta’s AI recognition algorithm.
Many photographers claim that simply changing the picture does not classify it as AI-generated. The fear is that this branding may mislead viewers or give the impression that their work is false.
Meta is currently building its AI detection technology, and it appears that they are experiencing some growing pains. Hopefully, they can improve the technique to reliably detect AI-created photos while keeping ordinary alterations out of the equation.
Photographers are divided. Some support transparency and believe any AI involvement, even basic editing tools, should be disclosed. Others argue that simple edits shouldn’t be confused with creating an entire image using AI.
Currently, Meta offers a single “Made with AI” label, regardless of the extent of AI use. For viewers, it’s unclear if a photo is a minor touch-up or a full-blown AI creation. Tapping the label reveals more details, but that requires extra effort.
Meta needs to improve their technology to differentiate between simple editing and full AI picture production. This ensures clear communication and avoids mistakenly devaluing photographers’ work.
