Meta is adding Made by AI label to real photos, photographers complain: Report

Meta is adding Made by AI label to real photos, photographers complain: Report

In March 2023, viral images of Pope Francis wearing a white puffer jacket and shades got people talking across the globe. It was soon discovered that these images were deepfakes and people were left perplexed. After all, how could an image that looked so real turn out to be AI-generated? Now, after a year, we all know that it is quite possible. And companies have started trying to find ways to fight this problem. One such company is Meta, which started labelling AI-generated images last month. But, it hasn't been without challenges. Photographers are now complaining that even the images clicked by them are being labelled as AI-generated by the company.

According to a TechCrunch report, the Facebook and Instagram parent company is facing backlash from users and photographers for tagging even real photos as "Made with AI." This issue has sparked widespread frustration, with many claiming that their human-made or minimally edited images are being incorrectly labelled by the platform's new AI detection system. In February, Meta announced its intention to label photos created with AI tools across its social networks. This move aimed to help users distinguish AI-generated content from genuine photos. Since May, Meta has been regularly tagging some images with the "Made with AI" label on its Facebook, Instagram, and Threads apps. However, this initiative is now under scanner for facing criticism.

Photographers and users have reported numerous instances of their images being mislabelled. A notable example, as highlighted by TechCrunch, is a photo of the Kolkata Knight Riders celebrating their Indian Premier League Cricket tournament win, which was incorrectly tagged as Made with AI. The label is only visible on the mobile apps and not on the web, adding to the confusion.

Pete Souza, a former White House photographer, experienced this issue firsthand. In a conversation with TechCrunch, he revealed that one of his photos was tagged with the new label, despite being a digital scan of a film image from the 1984 NBA Finals. Souza added that the need to "flatten the image" before saving it as a JPEG in Adobe's software might have triggered Meta's algorithm to attach the AI label. He expressed his frustration, stating that he was forced to include the tag even after unchecking it.

The TechCrunch report says that other users have encountered similar problems. A cosplay artist shared an image of their project which was incorrectly labelled as AI-generated. The artist revealed that the only editing done was for colour correction by the photographer. Another artist, who manually created pixel-based digital art, also reported their work being mislabelled.

Meta has acknowledged the feedback and is evaluating its approach to ensure the labels accurately reflect the amount of AI used in an image. A Meta spokesperson told the publication, "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."