The recent update to Meta’s AI labels was prompted by a wave of complaints from photographers who discovered that their authentic photos were being mistakenly labeled as AI-generated. The company has now changed the wording on these labels to “AI info” to address the confusion.
In early 2021, Meta introduced the “Made with AI” labels after reviewing their policy by the US Internet Regulatory Board. The board found that it was difficult to distinguish between images created with artificial intelligence and authentic ones, leading to some confusion for users. However, it quickly became clear that authentic photos were being mislabeled as AI-generated.
Upon further investigation, Meta discovered that the labeling was linked to images that had been retouched or enhanced using AI tools from various software. The company’s computer systems identified these manipulations and automatically applied AI labels, causing photographers like Pete Souza to raise concerns about the accuracy of these tags.
Souza, a former White House photographer, found that one of his photos was incorrectly tagged with the artificial intelligence label despite his efforts to remove it. He suspected that editing in Adobe’s Photoshop software triggered Meta’s algorithm to add the tag, even though he tried to remove it.
This week, Meta announced the correction of the label to “AI info” for images identified as being made with artificial intelligence. Users can now click on this label for more information about what constitutes an image with artificial intelligence characteristics and how Meta detects them using its system.
The change in wording on these labels is likely a response to concerns raised by photographers and other users who felt that they were being unfairly labeled as having used artificial intelligence when they hadn’t done so.
Overall, this update should help clear up any confusion about what constitutes an image created with artificial intelligence and how they are labeled on social media platforms like Meta.
