Facebook and Instagram users in the EU will soon be asked to share their public data for AI training. Meta strongly asserts that this time it will comply with privacy legislation.
Meta is starting to train its generative AI models based on public interactions and content from adult users in the EU. The company itself announces this in a blog post. Users will have the option to object to this use of their data.
What Does Meta Use?
Meta emphasizes that it will only feed public data from Facebook and Instagram users into its AI systems. Specifically, this includes public posts and comments you share via the company’s platforms. Interactions with Meta’s AI assistant, such as questions asked or comments made, will also be included in the training process.
Meta promises that it will stay away from your private conversations in Messenger and WhatsApp. Posts from minors also remain off-limits: they have already been seeing violent videos in their feed since Meta stopped the moderation policy with Donald Trump’s approval.
European Chatbot
The timing of the announcement is not coincidental. Last month, Meta AI was officially launched in Europe. The AI chat function has since been available in various messaging apps of the company. According to Meta, sharing European usage data should ensure that the chatbot will be better able to respond to the language, culture, and habits of European users.
Users from the EU will receive notifications in the apps and via email starting this week. These will explain which data Meta wants to use and why. The notifications also contain a link to a form through which users can file objections. Previously submitted objections remain valid.
Delicate Line
Meta is confident that this time it is in line with European privacy legislation. A year ago, Meta encountered resistance from privacy activists, causing it to pause training and launching AI products in Europe. The Irish and European privacy regulators have now given Meta the green light for its approach.
Meta even claims that it is now much more transparent about how it trains AI models than the other major American tech companies. It certainly can’t do worse than X. Elon Musk received a rebuke from regulators because X started collecting user data without asking first. The line between what is public and what is private is very thin in the context of AI training.
read also