Meta announced earlier this month that it would start labeling AI-created photos on its social networks. Meta has regularly tagged some photos with a “Made with AI” label on its Facebook, Instagram, and Threads apps since May. However, the approach has upset some photographers as real photos without AI are also being labeled.
There are examples of real photos that were taken and have the “Made with AI” tag, like one of a cricket team winning. The label only appears on mobile too. Other photographers also said real photos of theirs were wrongly flagged. Former White House photographer Pete Souza found a photo of his tagged as AI-made too.
Meta aims to label AI images using photo data. But simply editing a photo should not result in the label. Some tools like cropping or removing objects seem to falsely trigger Meta’s system. Photographers want clearer rules on what earns a photo the AI warning.
While some feel any AI use deserves disclosure, others think clean-up tools alone should not count. Users may also get confused as to how much AI altered a photo if the label appears too often. Meta has not commented on the mistakes or explained its labeling process well.
As elections near and deepfakes remain a risk, social media transparency on AI is important. But Meta’s new tagging system needs improvement to avoid incorrectly flagging real photographer’s works and confusing people.
