Meta to start labelling AI-generated content from May

Company to tag videos and images on Facebook and Instagram over the risk of deepfakes

Facebook and Instagram giant, Meta on Friday said it will begin labelling AI-generated content beginning this May.

The company tries to reassure users and government over the risks of deepfakes. The social media juggernaut added that it will no longer remove manipulated images and audio that don't otherwise break its rules, relying instead on labeling and contextualisation, so as to not unnecessarily infringe upon freedom of speech.

Tech insider websites such as The Verge also reported in February that Meta will start to label AI-generated photos uploaded to Facebook, Instagram and Threads as the world to observe elections in many countries. According the Nick Clegg, Meta’s president of global affairs, said in an interview that machine generated content is increasingly difficult to discern from reality. On the other hand, the White House also pushed the companies to watermark the AI-generated images and videos. However, Clegg assured that Meta is developing the tools to detect synthetic images and videos.

Meta is already adding watermark “Imagined with AI” to the images created with its own Imagine AI generator and will further expands this feature to various tools notably from Google, OpenAI, Microsoft, and Midjourney among others.

In a blog posted in February 2024, Meta asked people to consider several things about the content generated by AI, like trust worthiness of account and unnaturalness as well.