Meta to tag AI‑generated content on social platforms in new move

Starting in May, Facebook, Instagram, and Threads will introduce "made with AI" labels. These labels will identify images, videos, and audio materials created using artificial intelligence.

The swift advancement of artificial intelligence (AI) has recently heightened the divide between digital and analogue realities. The internet is now awash with texts, images, videos, and audio content crafted by algorithms. With the ongoing technological improvements, distinguishing between AI-generated content and content created by humans is becoming more challenging.

To help users navigate this new reality, Meta has revealed plans to start labelling AI-generated materials in May 2024. The initiative, spearheaded by the digital powerhouse owned by Mark Zuckerberg, will see video, audio, and image content tagged as "Made with AI". This action will be based on detecting AI indicators using Meta's algorithms and disclosures by individuals posting AI-generated content.

A survey conducted by Meta, involving more than 23,000 participants across 13 countries, aimed to gauge public opinion on how social media platforms should handle artificially generated content. According to findings published on the company's website, "A large majority (82%) favour warning labels for AI-generated content that depicts people saying things they did not say".

The introduction of labels for AI-generated content aims to combat misinformation and the phenomenon known as "model collapse". This term, which might sound overly dramatic, refers to programs that simulate human-generated content based on data from the internet. There's a concern that these programs might use their generated content as a data source, leading to a significant decline in the quality of new images, videos, etc., created. This scenario could render the essence of "artificial intelligence", which thrives on digitally altering human creations, obsolete.