Meta has officially launched its AI assistant, Meta AI, across the European Union, concluding a nearly year-long delay attributed to compliance with EU data protection regulations. This rollout includes integration into Meta’s suite of applications—WhatsApp, Instagram, Facebook, and Messenger—as well as availability via the web.
The current European version of Meta AI is limited to text-based functionalities, excluding features like image generation and multimodal interactions that are present in other regions. This limitation stems from Meta’s decision to withhold its most advanced Llama 4 models from the EU market due to regulatory concerns. The Llama 4 license explicitly prohibits use by entities domiciled in EU member states, reflecting Meta’s cautious approach to the EU’s stringent data protection landscape.
In terms of data usage, Meta plans to commence training its AI models using public content—such as posts and comments—shared by adult users on its platforms within the EU. This initiative, set to begin on May 27, 2025, excludes private messages and data from users under 18. Users will receive notifications detailing the data usage and will have the option to opt out through an objection form.
Despite these measures, concerns persist among privacy advocates and regulators. The Dutch privacy watchdog has expressed apprehension regarding Meta’s data usage practices, questioning the adequacy of the opt-out model in meeting legal requirements.
Meta’s strategic rollout in the EU reflects a balance between expanding its AI capabilities and adhering to regional data protection standards. While the current offerings are limited compared to other markets, Meta has indicated plans to enhance Meta AI’s functionalities in Europe over time, contingent upon ongoing regulatory engagements.
