The use of a new method in categorizing images by Meta has led to dissatisfaction among users and photographers, as a “Made with AI” label has been placed on images that were not created using artificial intelligence.
The company announced earlier in February its plan to classify images created using artificial intelligence on its social platforms.
Since May, Meta has been regularly labeling some images posted on Facebook, Instagram, and Threads as “Made with AI.”
There are numerous cases in which Meta automatically adds the “Made with AI” label to images that were not created using artificial intelligence.
Many photographers expressed concern about wrongly labeling their images as “Made with AI,” arguing that using photo editing tools should not result in such classification.
Pete Souza, former White House photographer, mentioned in a post on Instagram that one of his images includes the new label.
Souza pointed out that Adobe has altered the way the cropping tool works, suggesting users edit the image before saving it as a JPEG. He believes this change may be causing Meta’s algorithmic label to be attached to the image.
Souza stated that what bothered him was being required to add the “Made with AI” tag to his post, even though he had unchecked that option.
In a post in February, Meta stated that they rely on image metadata to determine the label.
At that time, the company announced, “We are developing tools capable of detecting invisible tags on a wide scale, allowing us to classify images from Google, Microsoft, Adobe, OpenAI, Midjourney, and Shutterstock as these companies implement plans to add metadata to the images created by their tools.”
It appears that Meta labels “Made with AI” when photographers use tools like Adobe Generative AI Fill to remove elements.
Meta has not specified when it will start applying the label automatically, although some photographers have supported its policy, considering it necessary to disclose any use of AI tools.
Currently, Meta does not provide specific classifications indicating whether a photographer used an editing tool or artificial intelligence to generate the image.
Users may find it difficult to assess the extent of AI usage in the image, and Meta explains that clicking on the label will display a message saying, “Machine-generated AI may have been used to create or modify content in this post.”
