Meta reverts to AI training on Facebook and Instagram posts despite legal pushback


Tech giant Meta plans to restart training its AI on EU citizens’ social media posts, the company announced on Monday, claiming the European Data Protection Board (EDPB) has “affirmed” the legality of its approach.

Meta halted its AI rollout in June following “intense engagement” with the Irish Data Protection Commission (IDPC) and complaints filed with data privacy watchdogs, known as DPAs, all over Europe.

The company recently launched text-only AI features in Europe, and, on Monday, announced a new opt-out mechanism which it claims legally allows Facebook and Instagram content to be used to train its AI models.

Beginning this week, EU users will receive a notification “to explain the kind of data we’re going to start using,” which will link to an easy-to-find form allowing users to object to their data being used, Meta said in a blog post.

“We welcome the opinion provided by the EDPB in December, which endorsed compliance in line with our original approach,” a Meta spokesperson told Euractiv.

The referenced EDPB opinion says that AI training on personal data could potentially be legal under a flexible GDPR exception, which allows a data processor’s “legitimate interest” to trump a data subject’s rights not to have their data processed.

“We are confident in our compliance with all data protection requirements and have engaged extensively with the IDPC on this matter,” the spokesperson said.

On Friday, the IDPC launched an investigation into Elon Musk’s X for use of users’ posts to AI training.

“Recommendations made by the DPC have led to a number of changes,” for example on user transparency, said an IDPC spokesperson, adding that they’ve been engaging with Meta since summer 2024 on its AI product.

“The DPC will continue to monitor this product as it rolls out,” the spokesperson said.

The practice will be challenged by privacy activists. The NGO Noyb filed 11 complaints with Meta the last time they started training on user posts, and a Noyb spokesperson said the practice is still unlawful, citing the need for prior user consent. 

The spokesperson disputed Meta’s claims about the EDPB opinion. It “simply stated that if you follow the law, the processing of personal data is lawful. But “European users’ fundamental right to data protection overrides Meta’s ‘legitimate interest’ for this extremely broad processing activity,” they said. 

“We have already filed complaints against Meta for this unlawful processing, and will be following up with the supervisory authorities,” the Noyb spokesperson added. 

(aw)



Source link

Share

Latest Updates

Frequently Asked Questions

Related Articles

YouTube to Revise Monetisation Policy to Target Mass-Produced and Repetitive Content

YouTube is updating its monetisation rules to increase its scrutiny of mass-produced content....

People Are Taking Massive Doses of Psychedelic Drugs and Using AI as a Tripsitter

Artificial intelligence, which is already trippy enough, has taken on a startling new...

Content moderators for Big Tech unite to tackle mental trauma

Content moderators from the Philippines to Turkey are uniting to push for greater...

Tesla’s German car sales continue to fall in June

Tesla's sales volume in Germany fell by 60% in June, the German road...
testing11