Meta AI Returns to Europe After Nearly a Year, Initially with Text-Based Chat

AIbase基地

Published inAI News · 4 min read · Mar 20, 2025

After nearly a year-long pause, Meta has announced the return of its AI chatbot, Meta AI, to the European market. Starting this week, Meta AI will be available across Meta's four major platforms – WhatsApp, Facebook, Instagram, and Messenger – in 41 European countries and 21 overseas territories. Initially, however, it will only offer text-based chat functionality.

Meta AI initially launched in the US in 2023. While Meta had planned to introduce the assistant to Europe, it was forced to postpone its rollout due to concerns raised by Ireland's data protection authority regarding the use of data from Facebook and Instagram users for model training. Similar regulatory concerns also led to the suspension of the rollout of its multi-modal Llama AI model in the EU.

Meta, Metaverse, Facebook

Currently, Meta says its AI assistant in Europe will function solely as a chatbot, primarily assisting users with brainstorming, trip planning, or answering specific questions using web information. European users can also use Meta AI to suggest content for their Instagram Stories. However, features like image generation, editing, or asking questions about photos are not yet available. Notably, the model has not been trained on the personal data of EU users.

Meta spokesperson Ellie Heatrick told The Verge, “After nearly a year of close engagement, we’re finally launching the service. For now, we are only offering a text-only model in the region, which has not been trained on first-party data from EU users. We will continue to work with regulators to bring the benefits of Meta’s AI innovations to the people of Europe and enjoy the advantages already available in other parts of the world.”

Last November, Meta began integrating some AI features into its Ray-Ban smart glasses in the EU, but these glasses currently lack multi-modal capabilities, preventing users from querying Meta AI about what they see. Meta doesn't appear to have abandoned plans to introduce more features to the European version of Meta AI, stating it will strive to “stay aligned with the US and expand our offerings over time.” The relaunch of Meta AI in Europe marks a significant step in its global rollout of AI assistants, while also demonstrating a cautious approach to navigating the complex regulatory landscape of Europe.

Meta AI Releases New Video Learning Model V-JEPA: A Breakthrough in Video Understanding

Recently, the Meta AI team launched the video joint embedding prediction architecture (V-JEPA) model, an innovative initiative aimed at advancing machine intelligence. Humans can naturally process information from visual signals and recognize surrounding objects and motion patterns. An important goal of machine learning is to reveal the fundamental principles that drive unsupervised learning in humans. Researchers proposed a key hypothesis—the predictive feature principle—arguing that the representations of continuous sensory inputs should be able to predict each other. Early research methods utilized slow feature analysis.

Meta Accused of Copyright Infringement: Using LibGen Dataset to Train AI and Deleting Copyright Information

Meta is facing a copyright infringement lawsuit, with the plaintiff's attorney claiming that CEO Mark Zuckerberg approved the use of a dataset containing pirated e-books and articles to train its Llama AI model. This case is one of many copyright lawsuits against several tech giants accused of using copyrighted works for AI model training without authorization. In documents filed in the U.S. District Court for the Northern District of California on Wednesday evening, the plaintiffs reiterated Meta's actions from the end of last year.

Meta Proposes a Novel Scalable Memory Layer to Enhance Language Model Knowledge and Reduce Hallucination Phenomena

As businesses increasingly adopt large language models (LLMs), improving the accuracy of model knowledge and reducing hallucination phenomena has become a significant challenge. Researchers at Meta AI have introduced a 'Scalable Memory Layer' in a recent paper, which may provide a solution to this issue. The core idea of the scalable memory layer is to add more parameters to the LLMs without increasing computational resources during inference, thereby enhancing their learning capabilities. This architecture is suitable for storing large amounts of factual knowledge while maintaining...

Meta to Launch AI Virtual Characters, Ushering in a New Era of 'AI Companionship' on Social Media

Social media giant Meta is brewing a significant transformation. According to the Financial Times, Meta's Vice President of Products, Conor Hayes, revealed that the company is actively developing an AI-generated character system that will play a crucial role on social platforms like Facebook in the future. Since launching its AI character creation tool in July last year, Meta has successfully incubated hundreds of thousands of AI characters. Although these virtual characters are currently in a private state, they will gradually make their way to the forefront of social platforms. According to Hayes, these AI characters will...

Facebook Plans to Introduce a Large Number of AI Virtual Users on Its Platform

Recently, Meta announced its plan to introduce a large number of AI-generated virtual characters on the Facebook platform, which will exist just like real users. Connor Hayes, Vice President of Product at Meta, stated that these AI characters will have personal profiles and avatars on the platform and will be able to generate and share content, with expectations of gradual adoption in the future. Meta's AI Studio has already launched the creation features for these virtual characters, allowing users to create them through the platform.