Key Points:

- Meta will begin using social media posts to train its AI models, but European users have the option to opt out, while users in other regions do not have this convenience.

- EU users can choose to refuse their public posts on Meta platforms to be used for training AI models, and this option will take effect globally on June 26.

- Meta's move may trigger opposition and challenges from European users and regulatory bodies.

ChinaZ.com, June 11th: Social media giant Meta will start using social media posts to train its AI models, although European users can opt out, users in other regions do not have this convenience. According to Meta's announcement today, this measure aims to provide more training data for its AI systems and has already begun to cover European users.

According to Meta, to better serve the European community, Meta's AI models need to be trained using information that reflects the diverse languages, geographical, and cultural backgrounds of the European population. Therefore, they hope to use the publicly shared posts from EU users on Meta platforms to train their AI models, including posts, comments, photos, and other public content, but private messages will be strictly prohibited from being used as training data.

However, Meta has sent billions of notifications to European users, offering them the opportunity to refuse before the AI training rules take effect globally on June 26. Meta states that any user can choose to refuse, and their posts will not be used for training AI models. Users outside the EU on Facebook and Instagram do not have this choice, and the training of future models will be mandatory for them.

Although Meta may believe it has sufficient reasons to start using European user data, it is hard to imagine there will be no opposition. Before Meta announced this news, consumer privacy advocacy group noyb had already filed lawsuits across Europe. noyb argues that the collection of user data should default to opt-in rather than opt-out. Additionally, due to the EU's Right to Be Forgotten, data is actually difficult to erase from large language models or other AI models, which could also pose problems.

Furthermore, the relationship between Meta and the EU is not very amicable. This year, the EU has launched several investigations into Meta, including investigations involving child safety and false information during the EU parliamentary elections. Although it is unclear whether Meta will ultimately be able to proceed with its plans, it is conceivable that at some point, the social network may face challenges.