Meta turns its users into AI "trainers"

Meta has announced that it will begin using the public content of European users to train its generative artificial intelligence models, including interactions with the Meta AI chatbot. The company has specified, however, that users in the European Union will be able to refuse the use of this data, through a form that will be distributed in the coming period.

• What data will be used?

According to the company's press release, the targeted data includes: comments and descriptions of photos; public posts on Facebook, Instagram, Messenger and WhatsApp; the content of conversations between users and Meta's AI systems.

The data will not include private messages between users, nor content generated by minors, the company claims. These clarifications come in the context of increased attention at the European level regarding the protection of online privacy.

• Notifications and opt-out

Users will be notified in the coming days about this policy change and will receive a link to the opt-out form. This way, each user will be able to decide whether or not they want their public activity to be used to improve Meta's AI systems.

• A delayed step due to European legislation

Meta AI was delayed in the European Union, precisely because of strict regulations on privacy and new technologies - in particular the General Data Protection Regulation (GDPR) and legislation on digital markets and artificial intelligence. The American company is now basing its approach on a decision of the European Data Protection Board in December, which stated that "legitimate interest" can be invoked as a legal basis for the use of certain personal data in the process of training AI models.