Earlier in February, Meta said it would start tagging photos created with AI tools on its social networks. Since May, Meta has regularly tagged some photos with the “Made with AI” tag on its Facebook, Instagram and Threads apps.
But the company’s approach to photo labeling has drawn the ire of users and photographers after it affixed the “Made with AI” label to photos that weren’t created using AI tools.
There are many examples of Meta automatically tagging non-AI created photos. For example, this photo of Kolkata Knight Riders winning the Indian Premier League cricket tournament. It should be noted that the label is only visible in mobile apps, not on the web.
Many other photographers have expressed concern that their images have been incorrectly tagged with the “Made with AI” label. Their view is that simply editing a photo with a tool should not be subject to etiquette.
Former White House photographer Pete Souza said in an Instagram post that one of his photos was tagged with the new tag. Souza told TechCrunch in an email that Adobe has changed the way its crop tool works and you must “flatten the image” before saving it as a JPEG image. He suspects that this action triggered Meta’s algorithm to attach that tag.
“What’s annoying is that the post forced me to include ‘Made with AI’ even though I unchecked it,” Souza told TechCrunch.
Meta did not officially respond to TechCrunch’s questions about Souza’s experience or the posts of other photographers who said their posts were incorrectly flagged. However, after the story was published, Meta said the company is evaluating its approach to indicate that the tags reflect the amount of AI used in the image.
“Our intention has always been to help people understand when they see content made with AI. We’re taking recent feedback into account and continuing to evaluate our approach so that our tags reflect the amount of AI used in the image,” a Meta spokesperson told TechCrunch.
In a blog post in February, Meta said it uses image metadata to detect the tag.
“We’re building industry-leading tools that can identify invisible tags at scale — specifically the ‘AI-generated’ information in the C2PA and IPTC technical standards — so we can tag images from Google, OpenAI, Microsoft, Adobe, Midjourney and Shutterstock as they implement their plans to add metadata to images created by their tools,” the company said at the time.
As PetaPixel reported last week, Meta appears to apply the “Made with AI” label when photographers use tools like Adobe’s Generative AI Fill to remove objects.
Although Meta did not clarify when it automatically applies the tag, some photographers have sided with Meta’s approach, arguing that any use of AI tools should be disclosed. The company told TechCrunch that it is actively working with companies that have AI-powered authoring tools to refine its approach.
“We rely on industry standard indicators that other companies include in content from their tools, so we actively work with those companies to improve the process so that our approach to labeling matches our intent.”
For now, Meta doesn’t provide separate labels to indicate whether a photographer used a tool to clean up their photo or used AI to create it. It can be difficult for users to understand how much AI is involved in the photo. The Meta tag specifies that “Generative AI may have been used to create or edit content in this post” — but only if you tap the tag.
Despite this approach, there are many photos on the Meta platforms that are clearly AI-generated and have not been flagged by the Meta algorithm. With the US election coming up in a few months, social media companies are under more pressure than ever to get AI-generated content right.
Story updated with Meta’s comments.
