Why is the deployment of Meta AI delayed in Europe?



Called to order by European, and particularly Irish, regulators, Meta must suspend the release, initially planned for this summer, of its AI assistant in Europe. At issue are the complaints from the noyb association, after Meta’s announcement of the use of its users’ data to train its language models.

European regulators force Meta to review its AI plans

Meta announced on Friday June 14, 2024, that it would postpone the release in Europe of its AI assistant integrated into Instagram, WhatsApp and Messenger. In fact, the personal data protection authorities have ordered Meta to review its deployment plans, but especially its training plans. The noyb association, of Austrian activist Max Schrems, filed eleven complaints in as many European Union countries after Meta’s announcement of the use of public content (photos, videos, comments, etc.) published by its users for training its AI models. The lack of clear consent, as well as a complicated process for refusing the use of one’s data, had motivated the association, customary in attacks against Meta.

So far, there has been no official change to Meta’s privacy policy, which would make this commitment legally binding, Max Schrems said.

The Irish regulator, in charge of the file, congratulated of the group’s announcement to suspend the release of its AI assistant in Europe, as well as “the training project using public content shared by adults on Facebook and Instagram in the European Union and the European Economic Area”. A decision that makes “following discussions with Meta”, specifies the regulator, which ensures that it is in constant discussion with the Menlo Park firm. The noyb association, for its part, does not intend to relax its vigilance: “We welcome this development, but we will monitor it closely (…) The complaints we have filed are ongoing and will require a decision”warned Max Schrems in a press release.

“Disappointed”Meta regrets “A step back”

Faced with requests from European regulators, Meta has therefore suspended its AI program in the European Union. In a press releaseMark Zuckerberg’s firm said “disappointed by the Irish Data Protection Commission’s request to delay the training of our large language models using public content shared on Facebook and Instagram”. And to add: “To put it simply, if we don’t include local information, we will only be able to give people a degraded experience. Which means we cannot launch Meta AI in Europe at the moment.”

This is a step backwards for European innovation, competition in AI development and further delays in bringing the benefits of AI to Europeans.

Meta remains convinced that its approach is “compliant with European laws and regulations”believing at the same time to be “more transparent than many of our industry peers”. Finally, he assures to continue working “so that Europeans have access to the same level of innovation in AI as the rest of the world”. For his part, noyb, still through the voice of Max Schrems, judges that Meta’s press release “looks a bit like collective punishment. If one European insists on his rights, the entire continent will not receive the new products. However, Meta has every option to deploy its AI based on valid consent. He simply chooses not to do it”.