Meta Secures EU Approval for AI Training with Social Media Data

Meta has received approval from the European Union's data regulator to utilize publicly shared content from its social media platforms to train its artificial intelligence models. The company stated that it will be using posts and comments from adult users on platforms like Facebook, Instagram, WhatsApp, and Messenger, as well as interactions with its AI assistant, to enhance its AI capabilities. Meta emphasized the necessity of training generative AI models on diverse data to capture the unique nuances and complexities of European communities, including local dialects and humor.

However, Meta has confirmed that private messages between users and public data from account holders under 18 years old will not be included in AI training. Users also have the option to opt out of having their data used for AI purposes through a forthcoming form that will be accessible in-app and via email.

Last year, Meta paused its AI training plans after privacy concerns were raised by a privacy advocacy group in multiple European countries, leading to a request from the Irish Data Protection Commission for a review. The complaints indicated that changes in Meta's privacy policies could allow extensive use of personal data for AI training. Now, with the EU's approval, Meta asserts that its AI training methods comply with legal requirements and maintains a constructive dialogue with the IDPC.

Meta noted that it follows practices similar to other tech companies, such as Google and OpenAI, which have previously utilized data from European users for their AI models.