Meta pulls plug on release of advanced AI model in EU

Mark Zuckerberg’s Meta will not release an advanced version of its artificial intelligence model in the EU, blaming the decision on the “unpredictable” behaviour of regulators.

The owner of Facebook, Instagram and WhatsApp is preparing to issue its Llama model in multimodal form, meaning it is able to work across text, video, images and audio instead of just one format. Llama is an open source model, allowing it to be freely downloaded and adapted by users.

However, a Meta spokesperson confirmed the model would not be available in the EU. The decision underlines tensions between big tech and Brussels amid a tougher regulatory environment.

“We will release a multimodal Llama model over the coming months – but not in the EU due to the unpredictable nature of the European regulatory environment,” the spokesperson said.

Brussels is introducing the EU AI Act, which comes into force next month, while new regulatory requirements are being put in place for big tech firms in the form of the Digital Markets Act (DMA).

However, Meta’s decision on the multimodal Llama model relates to whether it complies with GDPR (general data protection regulations). Meta has been ordered to stop training its AI models with posts from Facebook and Instagram users in the EU because of concerns it may violate privacy rules.

Ireland’s Data Protection Commission, which oversees Meta’s compliance with GDPR, said it was continuing discussions with the company over its model training.

However, Meta is concerned that other EU data watchdogs can intervene in the regulatory process and bring approval to a halt. Text-based versions of Llama are available in the EU and a new text-only version will be released in the EU soon – but those models were not trained on EU Meta user data.

The move follows Apple’s announcement last month that it will not roll out some of its new AI features in the EU owing to concerns about complying with the DMA.

Meta had been planning to use its multimodal Llama model in products such its Ray-Ban smart glasses and on smartphones. The Llama decision was first reported by Axios.

Meta also announced on Wednesday that it had suspended use of generative AI tools in Brazil after the government raised privacy concerns over the use of user data to train models. The company said it had decided to pause use of the tools while it held talks with Brazil’s data authority.