Meta has said it will withhold the roll out of future 'multimodal' artificial intelligence (AI) models in the EU over what it described as the "unpredictable nature" of the European regulatory environment.
Multimodal AI models are capable of processing data across video, audio, images and text, and across a variety of devices.
"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 parent company of Facebook, Instagram and WhatsApp said in a statement.
The company said that a text-only version of its Llama 3 model will be made available for customers and companies in the EU.
Last month, Meta announced that it was pausing plans to use personal data to train AI models after concerns were raised by the Irish Data Protection Commission (DPC).
Privacy campaigners had complained about Meta's plans amid fears that they may be in breach of EU privacy rules.
At the time, the DPC said it welcomed the decision by Meta to pause its plans to train its large language model using public content shared by adults on Facebook and Instagram across the EU/EEA.
"The DPC, in co-operation with its fellow EU data protection authorities, is continuing to engage with Meta on this issue," said Deputy Commissioner at the DPC Graham Doyle.
Advocacy group NOYB (None of Your Business) was among those who had urged national privacy watchdogs to act immediately to halt Meta's plans, saying recent changes in the company's privacy policy would allow it to use years of personal posts, private images or online tracking data for its AI technology.
Meta has said that competitors such as Google and OpenAI are already training on European data.
Last month, Meta said it was disappointed by the request from the DPC, its lead regulator, to delay training its large language models using public content shared by adults on Facebook and Instagram.
The company said it had incorporated regulatory feedback.
"This is a step backwards for European innovation, competition in AI development and further delays bringing the benefits of AI to people in Europe," Meta said at the time.
Meta's decision to withhold the roll out of new AI models in the EU was first reported by Axios.
