Meta has announced that it will begin training its artificial intelligence (AI) systems in the coming months using publicly available content shared by adult users on Facebook and Instagram in the UK.
“This means that our generative AI models will reflect British culture, history and idioms, and that UK companies and institutions will be able to use the latest technology,” the social media giant said. said.
As part of the process, users aged 18 and over are expected to receive in-app notifications starting this week on both Facebook and Instagram explaining how it works and how they can easily access the objection form , to opt out of using their data to train the company’s generative artificial intelligence models.
The company said it will respect users’ choices and that it will not contact users who already object to the use of their data for their purposes. It also notes that it will not include private messages with friends and family, as well as information from the accounts of minors.
In addition, Meta said the outcome was the result of its engagement with the UK Information Commissioner’s Office (ICO) and its recommendations supporting Meta’s compliance with the legitimate interest legal framework, which it said was a valid mechanism for using primary data to train its AI model.
“While our initial approach was more transparent than our industry peers, we have incorporated feedback from the ICO to make our objection form even simpler, more visible and easier to find,” added Meta.
It should be noted that Meta has suspended similar efforts in the European Union followed a request by the Irish Data Protection Commission (DPC) in June 2024. It called the move “a step backwards for European innovation”.
Austrian private non-profit organization Noyb since then accused the company shifts the burden to the users – ie. makes it an opt-out as opposed to an opt-in — and a failure to provide adequate information about how the company plans to use Facebook and Instagram’s publicly available data.
This comes after Meta suspended its use of generative artificial intelligence in Brazil after the country’s data protection authority issued a preliminary injunction objecting to a new privacy policy.
The ICO, in response to Meta’s plans, said it intends to monitor the situation when the company notifies users and starts processing their data.
“We were clear that any organization that uses its users’ information to train generative AI models needs to be transparent about how people’s data is used,” said Stephen Almond, the ICO’s executive director of regulatory risk. said.
“Organizations must put effective safeguards in place before they start using personal data to train models, including providing users with a clear and simple route to object to processing. The ICO has not given regulatory approval for the processing and Meta must ensure and demonstrate ongoing compliance.”
