Meta is relocating the AI info label in AI-edited or modified images
Meta has updated its AI labeling approach, distinguishing between AI-generated and AI-edited content by making the label less prominent for edited content and relocating it to the post's menu. However, concerns remain about the potential misinterpretation of heavily manipulated images.
In July, Meta announced it would modify its "Made with AI" label after reports surfaced that it had automatically attached itself to minimally retouched photos. To address the issue of the label possibly conflating AI-generated and heavily-modified media with minimally retouched content, Meta announced it would rename the label to "AI info". With this change, the company hoped to reduce the confusion, as the label would not imply that every marked piece of media was AI-generated, and users would be able to get the full context upon clicking on the label.
Now, Meta has announced another update which makes the AI label even less prominent in content that was only modified or edited using AI. In those cases, the label will no longer appear directly on the post and will be relocated to the post's menu instead. Content detected to have been AI-generated will still display the label, and Meta will share whether the label is due to self-disclosure or watermark-detection techniques. Meta claims that this update will better reflect the extent of AI used when creating content for its social media platforms.
However, even for AI-powered editing, there seems to be a wide range of outcomes, with using generative fill to change the aspect ratio of a real image being vastly different to modifying a picture to include objects or people that weren't there when the photo was taken. Treating those cases equally, with the AI info option hidden among the many options in the post's menu, may still lead some audiences to believe mistakenly that some AI-edited images may not be manipulated.
