• by Colin Kirkland , September 13, 2024

Following its Oversight Board's recent concerns over risking users' freedom of expression through the use of AI-editing tools, Meta has decided to make its once-clear labels on AI content edited or altered using AI tools less prominent across Facebook, Instagram and Threads.

This makes it more difficult for users to determine which images and videos are original and which have been modified.

Beginning next week, the company says, its “AI Info” tag will appear within a menu in the top-right corner of images and videos edited with AI, instead of directly beneath the user's name.

If they choose, users will have the option to click on the menu to check out the AI information, which should describe what about the image or video has been changed.

Meta's label changes only apply to images edited using AI tools, not images fully created using AI.

Any image generated entirely by AI will still be marked with an “AI Info” label directly on the post, explaining whether it has been edited by industry-shared signals, or if someone disclosed that it as an AI-generated image.

Meta says the changes are being enforced to “better reflect the extent of AI used” across images and videos on the platform -- a point of discussion within the company's Oversight Board, which in April called Meta's approach to AI-generated content was “too narrow” as it covered deepfakes, the most common manipulated media in 2020 when Meta's original AI content policy was written.

“In the last four years, and particularly in the last year, people have developed other kinds of realistic AI-generated content like audio and photos, and this technology is quickly evolving,” the company wrote in April. “As the Board noted, it's equally important to address manipulation that shows a person doing something they did not do.”

Meta has also stopped the immediate removal of all manipulated content, as the Oversight Board thinks the company should take a “less restrictive” approach to maintaining appropriate manipulated media. It seems these new labels are indicative of this strategy.

Meta began applying AI-generated content labels to videos, audio and images earlier this year, but after creators and photographers complained about their original work being flagged by mistake, the company updated its label from “Made with AI” to “AI Info.”

Despite the need for an updated approach to AI-generated content, Meta’s decision to limit the prominence of labels on manipulated media may further mislead users, especially as AI-editing technology becomes more convincing.