Meta to use public Facebook and Instagram posts to train AI

Meta has announced it will soon start training its artificial intelligence models using public Instagram and Facebook posts.

The tech giant said the process will start rolling out in the coming months after conversations with the UK’s data protection regulator, the Information Commissioner’s Office (ICO).

AI Meta told have not yet launched in the UK or EU due to concerns about the plans, but Meta has since said it had “engaged positively” with the ICO and said that the regulator had now given the scheme the green light.

“This clarity and certainty will help us bring AI at Meta products to the UK much sooner,” Meta said in a blog post.

“We welcome the ICO’s guidance supporting Meta’s implementation of the legal basis of ‘Legitimate Interests’, which can be a valid legal basis for using certain first-party data to train generative AI models for our AI at Meta features and experiences.”

The social media giant said it would not use people’s private messages with friends and family to train AI at Meta models, and that it does not use information from accounts of those in the UK under the age of 18.

It said it would use public information – public posts, comments, public photos and captions – from the accounts of adult users on Instagram and Facebook to “improve generative AI models for our AI at Meta features and experiences”.

The company said that from next week, UK users on the two platforms would start to receive in-app notifications to explain the scheme, as well as guidance on how to access an objection form where they can opt-out from their data being used to train the firm’s AI models.

Meta added that it had also incorporated feedback from the ICO to make its objection form “simpler, more prominent and easier to find”.

Stephen Almond, executive director of regulatory risk at the ICO, said it would “monitor the situation” in the coming weeks.

He added: “We have been clear that any organisation using its users’ information to train generative AI models to be transparent about how people’s data is being used.

“Organisations should put effective safeguards in place before they start using personal data for model training, including providing a clear and simple route for users to object to the processing. The ICO has not provided regulatory approval for the processing and it is for Meta to ensure and demonstrate ongoing compliance.”

Former deputy prime minister Sir Nick Clegg, now president of global affairs at the tech giant said the company was “pleased” to launch the scheme in the UK as well as Brazil.

In a post to X/Twitter, he said: “This is good news for innovation in both countries – it means we can bring our AI at Meta products to both the UK and Brazil much sooner, and that our generative AI models will understand local culture, history and idiom.

“Unfortunately, our plans to train our AI models to understand the EU’s rich cultural, social and historical contributions remain paused while EU regulators remain unable to agree on how the law should be applied.”

How to join the indy100’s free WhatsApp channel

Sign up for our free Indy100 weekly newsletter

Have your say in our news democracy. Click the upvote icon at the top of the page to help raise this article through the indy100 rankings.

Leave a Comment

url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url url