Meta makes a change to the labeling of posts containing AI-altered images on Instagram, Facebook, and Threads.
Meta announced a Changed labeling for posts containing AI-modified images in the apps of Instagram, Threads e Facebook. Starting next week, the notices will no longer be prominent, under the user’s name, but will be placed in a menu where users can find more information about the modified content. The announcement was made in an update to an April 2024 post on Meta’s official website. The post explains that the label “AI Info” (AI Information) will no longer be displayed under the user’s name in posts with AI-edited images and videos.
According to Meta, the change is intended to “better reflect the scale of AI being used” in the images and videos published on its platforms. The posts with images created entirely with AI will continue to have the label under the user’s name to prevent social media users from being misled.
It is possible that Meta has adopted this new strategy to avoid further criticism for labeling original content as AI-createdwhich had already been changed in July. In fact, there had been a lot of criticism from photographers and digital artists whose content was labeled as AI-generated even though it wasn’t, but had perhaps been simply modified with commonly used tools.
Meta did not specify what systems and metadata it uses to verify whether images have been modified with AI, limiting itself to saying that it uses “shared industry signals”like C2PA di Adobe and SynthID di Google.
Hiding labels like these in menus can make AI-modified content is difficult to identify on the Internet, which can make it easier for scams and disinformation campaigns to be implemented. Certainly, the various changes made by Meta reflect the difficulty in Finding the right balance in managing content modified or created with the help of Artificial Intelligence.