Meta Confirms It Will Train Its AI on Posts from European Facebook Users

The American digital giant Meta has confirmed that it plans to use posts from European users on Facebook to teach generative artificial intelligence, emphasizing that it is following the path paved by Google and OpenAI.

– We are following the example of others (companies), including Google and OpenAI – states in a statement from Meta Platforms released on Monday.

“Both Google and OpenAI have already used data from Europeans to train AI,” they emphasize at Meta.

– Our approach is more transparent and offers easier control than many of our counterparts in the sector who are already training their models based on similar publicly available information – adds Meta.

Meta announced that starting June 26, it will include content from users in Europe in the training of generative AI models. They plan to train their language model Llama based on content that citizens in the EU have publicly shared on platforms like Instagram and Facebook.

– Building this technology carries the responsibility of developing best practices and provisions that will comply with local laws and regulations. In line with this commitment, we are consulting with our main European privacy regulator – the Irish Data Protection Commission – and have incorporated their feedback into our training of Meta’s AI to ensure compliance with EU privacy laws – they state.

They note that since May 22, they have sent more than two billion notifications to users in Europe about the planned use of posts.

The Austrian association NOYB (None of Your Business) has requested from regulators in Austria, Belgium, France, Germany, Greece, Italy, Ireland, the Netherlands, Norway, Poland, and Spain to urgently stop Meta’s plans, emphasizing that the company should clearly ask users for consent to use their content.

Starting June 26, according to the planned changes, the company will be able to use posts and private photos of users and data about their online activity to teach AI technology, warn the Austrian digital rights association NOYB, founded by activist Max Schrems.

The process of denying consent is very complicated, emphasizes Schrems.

– Shifting the responsibility to the user is completely absurd. It is up to Meta to ask users for consent, as the law states, not to send them a well-hidden form for denying consent that leads to misleading conclusions – explains Schrems.

– If Meta wants to use your data, it must ask for your permission. Instead, they are forcing users to beg them not to use their data – said Schrems.

The head of the Austrian association dismisses Meta’s claim that its legitimate interest is to use data for training and developing models based on generative artificial intelligence and other similar tools.

– The EU Court resolved the issue of legitimate interest back in 2021 regarding advertising, and the company is now using the same arguments in matters of training AI models – warned Schrems.