Meta announced it is postponing the training of its large language models using public content from adult Facebook and Instagram users in the EU.
Meta announced it is delaying the training of its large language models (LLMs) using public content shared by adults on Facebook and Instagram following the Irish Data Protection Commission (DPC) request.
“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.” reads the DPC’s request. “The DPC, in co-operation with its fellow EU data protection authorities, will continue to engage with Meta on this issue.”
Meta added it is disappointed by request from the Irish Data Protection Commission (DPC), the social network giant pointed out that this is a step “backwards for European innovation, competition in AI development and further delays bringing the benefits of AI to people in Europe.”
“We’re disappointed by the request from the Irish Data Protection Commission (DPC), our lead regulator, on behalf of the European DPAs, to delay training our large language models (LLMs) using public content shared by adults on Facebook and Instagram — particularly since we incorporated regulatory feedback and the European DPAs have been informed since March.” reads the statement from Meta. “This is a step backwards for European innovation, competition in AI development and further delays bringing the benefits of AI to people in Europe.”
The company explained that its AI, including Llama LLM, is already available in other parts of the world. Meta explained that to provide a better service to its European communities, it needs to train the models on relevant information that reflects the diverse languages, geography and cultural references of the people in Europe. For this reason, the company initially planned to train its large language models using the content that its European users in the EU have publicly stated on its products and services.
Meta intended to implement these changes on June 26, giving users the option to opt out of data usage by submitting a request.
“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.” continues the statement.
“We are committed to bringing Meta AI, along with the models that power it, to more people around the world, including in Europe. But, put simply, without including local information we’d only be able to offer people a second-rate experience. This means we aren’t able to launch Meta AI in Europe at the moment.”
Meta added that the delay will allow it to address requests from the U.K. Information Commissioner’s Office (ICO) before starting the training.
Follow me on Twitter: @securityaffairs and Facebook and Mastodon
Pierluigi Paganini
(SecurityAffairs – hacking, Meta)