Meta has admitted to scraping all public posts made on Facebook and Instagram since 2007 to train its generative AI systems, sparking widespread privacy concerns. The revelation came during a public inquiry in Australia, but the practice applies globally.
While Meta had previously announced plans to use user data to train its AI, this is the first time the company confirmed it has been quietly doing so for over 17 years. Public posts, comments, and photos shared on these platforms have been used without explicit consent, raising questions about privacy, especially after Meta CEO Mark Zuckerberg’s claim that the company has more data than was used to train ChatGPT.
The company is legally required to offer opt-out options in the EU and UK, though investigations are underway regarding the legality of making it opt-out rather than opt-in. However, no such options were provided in other countries, including the US and Australia.
Meta’s global privacy director initially denied the data scraping practice but later admitted that unless users consciously set their posts to private, their public content has been used for AI training. Although Meta clarified that data from users under 18 was excluded, photos of children posted by their parents were not.
This revelation adds to the growing debate about privacy and AI, with many calling for more transparency and user control over their data.