Plans to use peoples’ public posts and images on Facebook and Instagram to train artificial intelligence (AI) tools belonging to parent company Meta have been attacked by digital rights groups.
The social media giant recently has been informing UK and European users of the platforms that, under privacy policy changes taking effect on June 26, their information can be used to “develop and improve” its AI products.
This includes posts, images, image captions, comments and Stories that users over the age of 18 have shared with a public audience on Facebook and Instagram, but not private messages.
Noyb, a European campaign group that advocates for digital rights, called its processing of years’ worth of user content on the sites an “abuse of personal data for AI”.
It has filed complaints with 11 data protection authorities across Europe, urging them to take immediate action on halt the company’s plans.
Meta said it was confident its approach complied with relevant privacy laws and was consistent with how other big tech firms used data to develop AI experiences across Europe. In a blogpost published on May 22, it said European user information would support a wider rollout of its generative AI experiences, in part by providing more relevant training data.
“These features and experiences need to be trained on information that reflects the diverse cultures and languages of the European communities,” it said.
Tech firms have been rushing to find fresh, multiformat data to build and improve models that can power chatbots, image generators and other buzzy AI products.
Meta chief executive Mark Zuckerberg said on an earnings call in February the firm’s “unique data” would be key to its AI “playbook” going forward.
“There are hundreds of billions of publicly shared images and tens of billions of public videos,” he told investors, also noting the firm’s access to an abundance of public text posts in comments.
The company’s chief product officer, Chris Cox, said in May the firm already uses public Facebook and Instagram user data for its generative AI products available elsewhere in the world. The way in which Meta has informed people about the change in the use of their data has also been criticised.
Facebook and Instagram users in the UK and Europe recently received a notification or email about how their information will be used for AI from 26 June.
This says the firm is relying on legitimate interests as its legal basis for processing their data – meaning people essentially have to opt-out by exercising their “right to object” if they do not want it to be used for AI.
Those wanting to do so can click the hyper-linked “right to object” text when opening the notification, which takes them to a form requiring they say how the processing would impact them.
The process has been criticised by Noyb, as well as people online who say they have tried to opt-out. In a series of posts about it on X, one user described it as “highly awkward”.
Another voiced concern that having to fill in a form and explain the processing’s impact on them might “dissuade” those who want to object from doing so. – BBC
