As of the end of June, Facebook was going to use all EU user data to train its language model, but backed out at the last minute.
The AI competition is in full swing, and although it's mostly about competing with Microsoft, OpenAI, and Google, Meta can't be left out of the list. at recent days Mark Zuckerberg It announced that AI functionality will soon appear on Facebook and Instagram, but for this it needs users' posts and photos as training data.
In order to do this, Meta shoved a simple consent form in users' faces instead of consulting, and only after a long questionnaire was there a chance to stop using our data. According to Meta, otherwise our data would need to reflect the services they provide “the different cultures and languages of European societies.”
For a long time, the company believed that it could obtain a supportive decision for data processing on the legal basis of “legitimate interests”, because this was necessary “for the development of Meta’s artificial intelligence”. However, that was not the case, she says Ars Technica.
Meta's EU regulator, the Irish Data Protection Commission (DPC), issued a statement confirming that following discussions with Meta about compliance with strict EU data protection laws, including the GDPR, the company has decided to suspend the processing of user data.
It also said it had suspended plans to train its large language model using public content shared by adults on Facebook and Instagram in the European Union and European Economic Area. Training was scheduled to officially begin on June 26.
