Meta has delayed plans to train its large language models (LLMs) using public Facebook and Instagram content posted adults in Europe following requests from regulators.

Last Monday, the social media giant announced its plans to expand its generative artificial intelligence features—which includes the Meta AI assistant—and the models underpinning them to Europe.

However, to “properly serve our European communities,” the company said that the LLMs “need to be trained on relevant information that reflects the diverse languages, geography and cultural references of the people in Europe who will use them.”

Adults’ public posts, public comments, and public photos and their captions would be used for training, the company said, and reiterated that peoples’ private messages with family and friends are not used to develop its AI systems.

Although the Big Tech company had been in consultation with—and incorporated feedback from—the Irish Data Protection Commission (DPC), its lead privacy regulator in the EU, the DPC has requested Meta to delay its training.

In an update confirming the delay, the social media giant said that it’s “disappointed by the request,” and that it’s a “step backwards for European innovation, competition in AI development and further delays bringing the benefits of AI to people in Europe.”

“We remain highly confident that our approach complies with European laws and regulations. AI training is not unique to our services, and we’re more transparent than many of our industry counterparts,” the company continued.

Recommended reading

• Ofcom: UK Hits Milestone in Broadband and Mobile Coverage
• Mobile UK Report Warns of 5G Risks Without Government Support
• BT Delays Plans for Forced Digital Landline Switch

As well as the DPC, the Information Commissioner’s Office (ICO)—the UK’s privacy regulator—also requested that Meta put a pause on the plans.

Meta said that specific requests from the ICO will also be addressed during this delay.

Speaking on Meta’s decision to halt training, Stephen Almond, executive director of regulatory risk at the ICO, commented: “We are pleased that Meta has reflected on the concerns we shared from users of their service in the UK, and responded to our request to pause and review plans to use Facebook and Instagram user data to train generative AI.

“In order to get the most out of generative AI and the opportunities it brings, it is crucial that the public can trust that their privacy rights will be respected from the outset.”

The DPC also remarked: “The DPC welcomes the decision by Meta to pause its plans to train its large language model using public content shared by adults on Facebook and Instagram across the EU/EEA.

“This decision followed intensive engagement between the DPC and Meta. The DPC, in co-operation with its fellow EU data protection authorities, will continue to engage with Meta on this issue.”