On Monday, Facebook owner Meta Platforms announced it would use European social media content to train its generative AI models
In a blog post, Meta announced that it will train its Llama large language models with content that individuals in the European Union have elected to share publicly on its platforms, including Facebook and Instagram.
The change aligns the company’s approach in Europe with how it handles the data it feeds into its AI models from other parts of the world, despite its previous apprehensions over the strict privacy and transparency regulations in the EU.
In a September interview with Reuters, Meta’s senior policy executive it was disclosed that the company employs public Facebook and Instagram posts to train its Llama models while excluding private posts and messages shared exclusively with friends.
Meta’s chief product officer informed Reuters that the company was “still working on the right way to do this in Europe” when the latest versions of Llama were released in April.
Last month, the social media behemoth announced that it would commence informing Facebook and Instagram users in the European region and the United Kingdom about how it utilizes public information shared on Meta’s services to enhance and develop AI.
In European countries, the advocacy group NYOB (none of your business) has lodged complaints challenging the move, claiming that the notifications were insufficient due to the EU privacy regulations, which mandated that Meta obtain user opt-in consent.
