Jun 15, 2024NewsroomArtificial Intelligence / Privacy
Meta announced on Friday that it is delaying its efforts in training the company’s large-language models ( LLMS ) using public content shared on Facebook and Instagram by adult users in the European Union , following a request made by the Irish Data Protection Commission.
The company expressed its disappointment at having to pause its AI plans, saying that it had taken feedback from regulators in the region and data protection authorities.
Meta’s plan of using personal data for training its artificial intelligence (AI), without asking users’ consent, is at issue. Instead, Meta relies on the legal basis’legitimate interests‘ to process first- and third-party information in the region.
The company had previously stated that users could opt-out from having their data used if they so desired. Meta already uses user-generated content in other markets, such as the U.S.
“This is a setback for European innovation and competition in AI development, and it further delays bringing AI benefits to people in Europe,” Stefano Fratta said , global engagement director at Meta privacy policy.
“We are confident that our approach is compliant with European laws and regulations.” Our services do not offer AI training exclusively, and we are more transparent than our industry peers.
It said that it could not bring Meta AI into Europe without the ability to train its AI model on local-collected data, which captures the varied languages, geographic references, and cultural references.
It noted that the delay would help it to address requests received from the U.K. regulator Information Commissioner’s Office, prior to beginning the training.
Stephen Almond, executive Director of Regulatory Risk at the ICO , said.
“We will continue monitoring major developers of generative AI including Meta to review the safeguards that they have put in to place and ensure that the information rights of U.K. Users are protected.”
The news comes after the Austrian non-profit organization noyb (none your business) has filed a complaint against 11 European countries, alleging that the General Data Protection Regulations (GDPR) were violated by collecting data from users to develop AI technologies without specifying them and sharing it with third parties.
Max Schrems, the founder of noyb, said that Meta is saying it can use “any data from any sources for any purpose” and make it accessible to anyone in any country, as long as they do it via “AI technology.” “This is the exact opposite of GDPR compliance.”
“Meta hasn’t said what it will do with the data, so it could be a simple bot, aggressive personalized advertising or even a killer robot. Meta also states that data from users can be shared with any ‘third party’, which means anyone around the world.
Noyb also criticized Meta’s disingenuous claims, and for framing the delays as a “collective penalty.” They pointed out that the GDPR privacy laws allows personal data to processed as long as the users give informed consent.
“Meta could roll out AI technology to Europe if it bothered to ask for consent, but it appears Meta is doing all it can to get opt-in consent to any processing,” said.