The company already faced data protection complaints in several EU countries.
ADVERTISEMENTUS tech giant Meta will not roll-out multimodel AI models – so-called virtual assistants – in Europe due to regulatory unpredictability, a company spokesperson confirmed to Euronews today (18 July).
“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,” a Meta spokesperson said.
The news, first reported by Axios, comes as the company had already paused the roll-out of its AI assistant in Europe after the Irish Data Protection Commission told Meta to postpone its plan to use data from Facebook and Instagram adult users to train large language models (LLMs).
Meta updated its privacy policy asking to take all public and non-public user data – with the exception of chats between individuals – for use in current and future AI technology, which was due to take effect on 26 June.
In response, Austrian privacy organisation NOYB filed complaints with privacy watchdogs in eleven EU member states, alleging Meta’s practices were not compliant with the EU’s General Data Protection Regulation (GDPR).
NOYB requested an «urgency procedure» under the EU’s data protection rules. It claims the change is worrying because it involves the personal data of about 4 billion Meta users.
Describing it as a “step backwards” for European innovation, Meta said at the time that it remained highly confident that its approach complies with European laws and regulations.
The Irish Data Protection Commission (DPC) told Euronews in June that “Meta delayed the launch following a number [of] enquiries from the DPC which have been addressed”. Meta gave users four weeks’ notice ahead of the initial training, the DPC said.
Meta has its own large language model called Llama, the latest version of which (Llama 3) was released in April and is used to power its assistant Meta AI.
