Meta’s artificial intelligence (AI) content detection tool on Instagram has been spotted tagging real images with the ‘Made by AI’ tag. One such post that was incorrectly flagged was from the official Instagram account of Kolkata Knight Riders who recently won the Indian Premier League (IPL) 2024 cricket tournament. One of the images posted on the account, showing the team lifting the trophy, was tagged by the platform as generated by artificial intelligence. Several photographers on the social media platform have experienced the same problem.
In February, Meta said it was in the process of introducing an AI-generated content detection feature that would prevent users from misinformation and highlight cases of deepfakes (AI-generated or digitally altered images and videos made to resemble another person, location, or event). The feature recently launched on Instagram, and it also appears to tag real photos as AI-generated content. Currently, these tags can only be seen in the iOS and Android apps, but not on the web.
Photo of Kolkata Knight Rider tagged ‘Made by AI’
Photo credit: Instagram/kkriders
While KKR’s photo is one of the most prominent examples of this mistake, users have called out many other such mislabels. Among them is former White House photographer Pete Souza, who posted a photo of an old basketball game.
After the wrong tag was added, he edited the caption to read: “I don’t understand why Instagram is using “made with AI” on my post. There is no artificial intelligence with my photos.” He also stressed that he was unable to undo the tag because the platform kept adding it.
Frustrated users also began flooding Threads, a text-based social media platform signed by Meta, tagging Instagram host Adam Mosseri to highlight the issue. One user said: “Not a single photographer or artist on all of Facebook and Instagram has any idea what triggers the “Made with AI” tag. Although Mosseri clarified that all they do is read C2PA tags, no one understands how to avoid it.”
Earlier, Meta’s president of global affairs, Nick Clegg, said the company was working with “industry partners to align with common technical standards that signal when a piece of content has been created using artificial intelligence.” It also claimed that the detection tool can correctly tag images from Google, OpenAI, Microsoft, Adobe, Midjourney and Shutterstock.
However, the implementation of the feature appears to be flawed. PetaPixel found in a report that even removing a tiny detail in an image using Adobe’s AI-powered Generative Fill gives images a ‘Made by AI’ label. However, non-AI tools such as the Spot Healing Brush or Clone Stamp tools did not add the mark despite the result being the same.
The publication also found that when an image previously labeled as AI-generated was loaded back into Photoshop and saved after copying and pasting onto a black document, the AI label did not appear.
Meta spokeswoman Kate McLaughlin told The Verge that the company is now taking recent user feedback into account and evaluating its approach. “We rely on industry-standard metrics that other companies include in the content of their tools, so we’re actively working with those companies to improve processes so that our tagging approach matches our intent,” McLaughlin was quoted as saying in the publication.
