Images that have been generated by artificial intelligence will become more identifiable in the future thanks to Meta. The group wants to be able, within months, to label this type of files shared on its networks as Facebook Or Instagram.
Generative AI is a particularly fascinating technology with its ability to produce text and images based on a request formulated in a few words. But in the wrong hands, it can also convey misleading information, which can have potentially catastrophic effects on societies. So, like others before it, the Meta group wants to be able to distinguish real images from fakes. And it should be possible by the end of next spring.
Meta wants to monitor AI-generated images
We are talking more and more about the need to label the images produced by AI, so as not to confuse them with reality. An area in which Meta will fully commit according to its president of global affairs, Nick Clegg. To do this, the company will use the presence of invisible markers present in this type of document, to detect them, and then publicly display their origin.
Users of Facebook, Instagram or Threads will be informed that an image was produced by artificial intelligence. Meta, which already imposes this labeling for all documents generated by its tools, wants to be able to repeat the operation on images produced using AI from other giants in the sector such as OpenAI, Microsoft, Midjourney or Google.
Images, the easiest content to detect
An initiative which will be welcome in this year 2024, which promises to be crucial at the global level with many elections planned, including the European elections, and especially at the end of the year, the American presidential elections. These ballots should therefore not be hampered by the potential massive irruption of false images trying to pass themselves off as real.
The fact remains that this is a project that is only just beginning. Because AI not only produces images, but also texts, audios and videos. As for the written word, it seems that it is already too late to be able to distinguish the human hand from the artificial origin. “ This train has passed »explained Nick Clegg. For audios and videos, there is more hope, even if at present no correct solution exists.
“ While the technology isn’t quite there yet, especially when it comes to audio and video, we hope we can build momentum and inspire the rest of the industry to follow » wants to reassure Nick Clegg.
Source : Reuters
0