Meta, the owner of Facebook and Instagram, has announced that it will start using social media content from Europe to train its generative AI models.
According to a blog post released yesterday, Meta will be training its Llama large language models using public content shared by people in the EU on its platforms.
This decision brings Meta’s approach in Europe more in line with its data handling practices in other parts of the world, despite previous caution due to strict EU privacy and transparency regulations.
An advocacy group called NYOB (none of your business) has filed complaints in various European countries, arguing that the notifications provided were inadequate and that Meta should have obtained opt-in consent from users as required by EU privacy rules.
Meanwhile, in the US, the Supreme Court is set to review a multibillion dollar shareholder lawsuit against Meta. The lawsuit accuses the social media giant of misleading investors about the Cambridge Analytica data-harvesting scandal.
The court will determine whether a federal appeals court made a mistake in allowing the lawsuit to proceed based on allegations that Facebook (as it was known at the time) inflated share prices by failing to adequately disclose the risk of misuse of user data.
