Meta has announced that it is resuming plans to train its AI models using public posts from UK Facebook and Instagram users. After pausing its efforts three months ago due to regulatory pressure, the tech giant claims that it has now “incorporated regulatory feedback” and improved its approach to be “more transparent.”
Please follow us on Facebook and Twitter.
Starting next week, UK users will see notifications in their apps, explaining how their public content could be used to train Meta’s generative AI models. Meta says its goal is to ensure that these AI systems reflect “British culture, history, and idiom.” Users will have the option to object to their data being used, but only if they actively opt-out through the process provided by Meta.
Revised Opt-Out Process
The initial suspension of Meta’s plans came after the UK’s Information Commissioner’s Office (ICO) raised concerns about the way Meta was handling user data for AI training. The Irish Data Protection Commission, Meta’s primary privacy regulator in the European Union (EU), had also objected to the plans, citing feedback from multiple EU data protection bodies.
This regulatory pressure has made it difficult for Meta to move forward with its AI training plans across Europe. While Meta has been using user-generated content to train its AI models in the United States, the more stringent privacy regulations in Europe have proven challenging for the company.
According to TechCrunch, Meta is now resuming its efforts in the UK, where data protection laws are still based on the EU’s General Data Protection Regulation (GDPR), but enforcement practices differ, especially after the UK left the EU. Despite Meta’s past reliance on “legitimate interest” (LI) as the legal basis for its data use, questions remain about whether this justification is appropriate under GDPR.
ICO Response
Meta’s renewed plans include a simplified opt-out process, though users will still need to go through a formal process to notify Meta that they do not want their data used for AI training. Those who had previously opted out will not need to resubmit their requests. Stephen Almond, the ICO’s director of technology and innovation, has stated that the ICO will “monitor the situation” as Meta moves forward.
Almond emphasized that Meta is responsible for ensuring its data practices comply with UK law, adding that companies using personal data for AI training must be clear about how that data is being used. He also called for easy-to-use opt-out mechanisms, urging organizations to follow established privacy guidelines.
Meta insists that its revised process is simpler than before but has yet to provide specific details on how it has improved the user experience. The company’s previous opt-out process involved multiple steps and required users to provide a reason for objecting, with Meta reserving the right to accept or deny those objections.
End Note
For now, Meta’s AI training in the EU remains on hold, with no clear timeline for when or if it will resume. Meanwhile, UK users are set to see the revised notifications in their apps soon, as Meta works to balance AI advancements with growing privacy concerns. How this new chapter in Meta’s AI journey will unfold remains to be seen, especially as privacy advocates and regulators continue to scrutinize the company’s data practices.