After nearly a year of pause due to regulatory issues, Meta has begun collecting public content from its European users to train its AI models, just as EU officials are preparing to issue the first-ever fine under the Bloc’s Digital Markets Act (DMA).
Meta announced on Monday that it will begin using public posts, comments and AI interactions from adult users of Facebook, Instagram and WhatsApp in the EU to improve the AI system it generates. The European Data Protection Commission (EDPB) approved the launch.
“This training will better support millions of people and businesses in Europe by teaching our generative AI models,” Mehta said in a formal announcement.
The use of EU data was previously prohibited. Explain 2024 “There is no EU user data, the metadata says: “We can only provide people with a second-rate experience. This means we are currently unable to launch Meta AI in Europe. ”
META provides exit options
European users will start receiving notifications in the apps and emails this week and explain exactly what data will be collected and how to use it. These notifications will include links to objection forms that users can opt out.
“We make this form of objection easy to find, read and use, and we will respect all objection forms we have received and newly submitted objection forms.” Message press release says.
“We do not use private information from people with friends and family to train our generative AI models,” the company stressed.
Furthermore, “public data from the accounts of EU people under the age of 18 are not used for training purposes.”
Mehta also noted that this is “an example of following settings from others, including Google and Openai”, noting that both companies have used data from European users to train their AI models.
DMA fine
The timing of the Meta announcement is worth noting, just as the European Commission is preparing to issue fines against the dollar and Apple for alleged violations of the New Digital Markets Act.
Competition Commissioner Teresa Ribera strengthened the commission’s enforcement intentions on Tuesday and told the European Parliament: “If we don’t see the willingness to cooperate, we will not evade the imposition of the fines determined by the law.”
Companies found to violate DMA can fine up to 10% of their global turnover, while repeated violations increase to 20%.
While the EU wants to fine large technologies, it turns out that preventing model training of citizen data is impossible.
For European users who do not want data to be harvested by metadata, please be aware of meta notifications.
