Users can reportedly change their profile settings and switch to private accounts to prevent Meta from using their data, Facebook, and Instagram posts to train its artificial intelligence models.
Making accounts private could be an intelligent workaround for Meta, which has been integrating public data from US accounts into its generative AI capabilities without providing an opt-out choice.
People's private information is used as training data for AI chatbots and other huge language models, like those that generate visuals. Businesses like Google and OpenAI use publicly available data to train their AI algorithms. It's crucial to remember that these businesses allegedly do not give AI users' social media data directly.
Rather, they depended on information shared by publications and tiny websites. Additionally, they reached deals to use content lawfully with major publishers and businesses, including News Corp and Reddit.
But Meta's AI instruction isn't the same. The firm will use every private aspect of a user's life that they disclose to the public. This applies to captions on a user's posts and Reels, as well as the images and videos in their feed. Only public accounts are eligible for this information to be used by Meta, sparing private accounts, Instagram and Facebook stories, and Threads data.
According to Meta, it also doesn't use any information from direct, private messages on Facebook or Instagram, even for users who have public accounts.
Privacy Against AI Training
Users can make their Facebook profile private by simply going to Facebook's "settings & privacy," tapping "settings," and then finding the "Audience and visibility" section. From there, users can click "posts" and then select the option "Friends" or "Only me" rather than "Public."
As for Instagram, users must head onto the profile tab, access the "three lines" at the top right corner, and then find "Settings and Activity." From there, users must access "Account Privacy" and make their accounts private.
Concerns about Meta's AI Training Data
Meta's recent announcement of training their artificial intelligence models using Facebook and Instagram posts from 2007 has prompted various regulatory bodies to voice their concerns with the new policy.
In 11 complaints submitted to the EU, the advocacy group NOYB (none of your business) claimed that Meta planned to use years' worth of personal images, posts, and internet tracking data for an unknown "AI technology" that could collect personal data from any source and share it with an unlimited number of unidentified third parties.
The organization is also appealing to authorities to reverse the policy. According to NOYB, about 400 million users across Europe would be impacted. The group expressed concern about the requirement that users consciously decide not to submit data in the future.
