Meta Withholds AI Models from EU Over Regulatory Uncertainty

Meta has decided to withhold its upcoming multimodal AI model and future models from customers in the European Union. This decision stems from what Meta describes as a lack of regulatory clarity in the region, Axios has learned.

Growing Tensions with EU Regulators

The move underscores increasing tensions between U.S. tech giants and EU regulators. "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," Meta stated to Axios. This decision echoes a similar stance by Apple, which recently withheld its Apple Intelligence features in Europe over regulatory concerns.

Implications for European Market

Meta's new multimodal models, capable of reasoning across video, audio, images, and text, will not be available to European companies, even though they will be released under an open license. This could also prevent non-EU companies from offering products and services in Europe using these models. However, Meta plans to release a larger, text-only version of its Llama 3 model soon, which will be available to EU customers.

Regulatory Challenges and GDPR

Meta's issue lies not with the impending AI Act but with compliance with GDPR when using European customer data for training models. In May, Meta announced its intention to use publicly available posts from Facebook and Instagram users for training, offering an opt-out mechanism to EU users. Despite briefing EU regulators in advance, Meta received minimal feedback initially but was later ordered to pause training and faced numerous questions from data privacy regulators across Europe.

Comparison with the UK

Interestingly, Meta does not face the same regulatory uncertainty in the UK, which has a law nearly identical to GDPR. Meta plans to launch its new model for UK users without similar concerns.

Impact on Tech and Regulatory Landscape

Meta's decision highlights the broader conflict between U.S.-based tech companies and European regulators, who are seen as more stringent on privacy and antitrust issues. Tech companies argue that such regulations harm both consumers and the competitiveness of European firms. Training on European data is crucial for Meta to ensure its products reflect the region's terminology and culture accurately. Meta pointed out that competitors like Google and OpenAI are already training on European data.