Meta said it would use public user posts from European users on its platform to train its AI.
On April 14, Meta announced that it would train its AI using public user content on its European platforms, including posts and comments from people over the age of 18, along with information they shared with the Meta AI artificial intelligence tool.
Users will receive a notification in the app or via email, and they can fill out a form to object if they wish. Meta explains that training AI with European data helps to better support millions of users and businesses in the region, as the AI will better understand and reflect their cultures, languages, and histories. "Our generative AI models are trained on a wide range of data to understand the incredible and diverse nuances and complexities that make up European communities," the company writes.
In fact, Mark Zuckerberg's company is using data from most of the world's users to train its AI. Some investigations have found that the data may have been collected in posts dating back to 2007. "We use public posts and comments on Facebook and Instagram to train our AI models," Meta confirmed last year.
In Europe alone, the company that owns social networks Facebook, Instagram, Threads planned to collect user data from June 2024, but had to postpone it due to regulations related to data protection and privacy in the region. The company now says it "met its legal obligations" before implementing.
According to experts, publicly collecting user data shows that companies may be "hungry" for data to train AI. Gizmodo calls Meta's approach feeding data into a "content grinder" and a way to implicitly declare that user content belongs to the platform.
"This shows that they need your content, not you," the site wrote.
Last weekend, Jack Dorsey, CEO of Block and founder of Twitter, also caused controversy when he called for "abolishing all intellectual property laws", which Elon Musk agreed with and shared. The call comes from the fact that AI models have difficulty supplementing training data, leading developers to want to obtain copyrighted data sources.
In the post, Meta also asserted: "We pride ourselves on being much more transparent in our approach than our competitors in the industry."
(according to Gizmodo, The Verge)