Meta, the owner of WhatsApp, Instagram, and Facebook, has decided to implement a labeling system to distinguish AI-generated images across its social networks in the coming months. This initiative aims to help differentiate between what is real and what is fake.
On Tuesday, Meta announced it is working with industry partners to establish technical standards that will facilitate the recognition of images, videos, and audio generated by AI tools. Analysts are closely monitoring how successful this experiment will be, especially as it has become extremely easy to create and disseminate images via AI that can cause harm, including misinformation in political contexts and creating fake nude images of celebrities without their consent.
Gilly Fidan, an assistant professor of information technology at Cornell University, commented, "This step signals that they are paying significant attention to the seriousness of generating fake content online, which poses a problem for their platforms." She added, "This step could be very effective in reporting a significant portion of the false content generated by AI, but it is unlikely to reveal everything."
Nick Clegg, Meta's president of international affairs, did not specify on Tuesday when these labels would start appearing, but he mentioned they will be available in multiple languages within the coming months, particularly in light of "a number of important elections happening around the world." He stated, "As the distinction between human and AI content becomes unclear, people want to know where the boundaries lie."
Meta began applying such labels on innovative images created using its "Meta AI" tool since its launch in December. This move comes at a time when the rise of AI raises concerns about individuals potentially using it to create political chaos through misinformation or spreading false information in the lead-up to several crucial elections taking place this year across multiple countries.
