Meta suspend plans to train AI on EU users’ data

In this post:

  • Meta has postponed its plans to train AI models on the social data of EU Facebook and Instagram users.
  • Irish regulators requested to halt its decision as it made the request on behalf of other EU regulators.
  • Meta has called the decision a step backward in European innovation as it says it will be able to serve EU users fairly.

Meta has seemingly halted plans to process vast amounts of user data to introduce new AI features in the European market. The decision came after regulators asked the company to pause its plan to train AI models on users’ posts.

Also read: Meta will use users’ social activities to train its AI models

Meta has now updated its previous statement about training its large language model (LLM), Llama, on user-generated content. The company said it postponed its plans after the Irish Data Protection Commission (DPC) objected to the proposal.

Meta had an interest in processing EU users’ data

On Monday, Facebook parent Meta announced that it would train its AI models on peoples’ public posts, images, image captions, and their conversations with the chatbot. However, the Irish regulator opposed Meta’s stance that it had “legitimate interests” in processing data of users belonging to users from the European Union and European Economic Area (EEA) for its AI models. The company was also planning to use data from Instagram. Meta noted in a blog post that,

“This is a step backward for European innovation, competition in AI development, and further delays bringing the benefits of AI to people in Europe.”

Only a little information is available, but Meta has confirmed that it has made the decision. Meta also said that the tech firm is disappointed with the request from the Irish regulator (DPC), who is the lead regulator on behalf of all the European data protection authorities.

Meta also said that a factor of disappointment comes from the fact that it incorporated regulatory feedback and informed the European authorities in March this year. The company shared that this delay will also enable it to address requests made by the Information Commission Office and UK regulator before it starts training its AI models.

EU regulators will engage with Meta on this issue

The Irish regulator also published a short note on its website stating that it welcomes Meta’s decision to pause its program to start training LLMs using content shared by citizens residing in the EU/EEA territory. The DPC said that,

“This decision followed intensive engagement between the DPC and Meta.” 

The regulator also showed its intentions of continuously engaging with Meta on this issue in cooperation with other EU data authorities. The Irish and other EU DPCs are not the only ones opposing Meta’s idea of using public content for AI training. 

Along with public backlash, the European Center for Digital Rights called Noyb also complained to data protection authorities in different EU countries. Noyb also said that regulators initially gave Meta the go-ahead to continue its plans, but they have now reversed their decision. 

Noyb filed complaints against Meta in Spain, Austria, Poland, France, Netherlands, Belgium, Greece, Germany, Norway, Ireland, and Italy. Noyb said that it will file more complaints to stop Meta from its plans to use public social data.

OpenAI and Google also used EU users data to train their AI

Meta previously announced that it will bring new AI features to EU users. The company planned to incorporate AI in almost every way people use their products, including stories and chats. Meta said that with training on EU data, its AI will be able to serve European customers with quality services. 

Meta argues that without EU data, its AI systems will not understand the diverse cultures and important languages. The company said that EU citizens will be “ill-served” by its AI models because they will not be trained on rich European cultural and social contributions.

Also read: Wall Street reacts as Meta announces ways it will be draining funds on AI projects

Meta also argued that there are other companies to follow this practice. It stated examples of OpenAI and Google, which have used EU users’ data to train their models. Meta said that their practices are more transparent and provide users more control. However, Noyb’s stand is that Meta’s practices did not comply with the General Data Protection Regulation (GDPR). 

Meta previously gave EU users a deadline of June 26 to opt out, while users in other areas, except the US state of Illinois, still do not have such an option. Noyb argued that handling the case this way will leave users with no option to opt out of the system in the future.

Cryptopolitan reporting by Aamir Sheikh

Subjects tagged in this post: | |

Disclaimer. The information provided is not trading advice. Cryptopolitan.com holds no liability for any investments made based on the information provided on this page. We strongly recommend independent research and/or consultation with a qualified professional before making any investment decisions.

Share link:

Most read

Loading Most Read articles...

Stay on top of crypto news, get daily updates in your inbox

Related News

Apple will bring AI to the Vision Pro
Subscribe to CryptoPolitan