Meta to stop offering future multimodal AI models in EU over 'lack of clarity' from regulators

Meta is pulling back on releasing its upcoming multimodal AI model, Llama, and any future models in the European Union. This decision, driven by what Meta calls regulatory uncertainty, shows a trend among US tech companies to withhold their products from the European market due to unclear regulations.

Meta’s new multimodal AI model can process video, audio, images, and text, aiming to enhance products like smartphones and Meta Ray-Ban smart glasses. However, this won’t be available in the EU. A statement from Meta mentions that the unpredictable regulatory environment in Europe is the reason for this decision. This follows Apple’s recent decision to delay its Apple Intelligence features in Europe over similar concerns. The Irish Data Protection Commission, Meta’s lead privacy regulator in Europe, hasn’t commented on this issue yet.

The implications are significant: European companies won’t have access to Meta’s advanced AI models, and companies outside the EU may also struggle to offer products in Europe that use these models. On the other hand, Meta plans to release a larger, text-only version of the Llama 3 model, which will be available to customers and companies in the EU.

Meta’s main challenge isn’t with the upcoming AI Act but with the General Data Protection Regulation (GDPR) on how to train models using data from European customers. In May, Meta announced plans to use publicly available Facebook and Instagram posts to train future models, notifying over 2 billion EU users about opting out. However, after publicly announcing these plans, Meta was ordered to pause the training on EU data and received numerous questions from data privacy regulators.

Interestingly, the United Kingdom, with regulations similar to GDPR, hasn’t posed the same level of uncertainty. Meta plans to launch its new model for UK users without issue. A Meta representative noted that European regulators are slower in interpreting existing laws compared to other regions.

This situation highlights the growing tension between US tech giants and European regulators. The EU is known for its strict privacy and antitrust regulations, which tech companies argue can stifle innovation and competitiveness. Meta stressed the importance of training on European data to ensure their products reflect regional terminology and culture accurately. They also pointed out that competitors like Google and OpenAI are already training on European data, adding to the competitive pressure.

Overall, Meta’s decision not to release its multimodal AI models in the EU underscores the challenges tech companies face with European regulations, and it sets the stage for ongoing conflicts between the tech industry and regulatory bodies.