Meta caves to Europe to launch its AI assistant
Meta has finally bowed to European demands. This week, Mark Zuckerberg’s company is rolling out its AI assistant across the continent, nine months after abandoning its initial plans due to a dispute with data protection authorities. To move forward, Meta had to scrap its original approach, which involved training its Llama language models using users’ messages and photos.
“It’s taken longer than we would have liked,” the company admitted, citing a “complex regulatory system.” However, this launch remains incomplete: some features, such as image generation, will not be available. Meta insists this is only a “first step” and aims to achieve “parity with the United States.”
Named Meta AI, the assistant is available on the company’s four main platforms: Facebook, Messenger, Instagram, and WhatsApp. Users can interact with it directly or within conversations with friends. The AI can answer questions, edit photos, and even translate videos automatically. Meta AI is powered by the latest multimodal Llama model, allowing it to process and generate text, audio, and images. Initially launched in the US a year and a half ago, the assistant has over 700 million active monthly users—more than ChatGPT.
“Meta to launch generative AI assistant in European Union”
According to CNBC, it will soon have its own dedicated app. And a premium version is on the way. Meta executives have also hinted at the future introduction of ads.
Challenges in the European Market
Last year, the company postponed the European launch of its chatbot due to opposition from Ireland’s Data Protection Commission (DPC), its primary regulator under the General Data Protection Regulation (GDPR). Pressured by its European counterparts, the DPC prohibited Meta from using users’ posts, comments, and photos to train its models. The reason? Meta had refused to seek users’ explicit consent, instead allowing them to opt out by filling out an online form—provided they gave an acceptable justification. Austrian privacy activist Max Schrems slammed this approach as “clearly in violation of the GDPR” and took legal action.
Adapting to European Regulations
Meta argued that without access to European users’ public content, it could offer only a “second-rate experience,” leading to the decision to cancel Meta AI’s rollout. “If we don’t train our models on public content that Europeans share on our services, the AI features won’t accurately understand important regional languages, cultures, or trending topics,” the company said—though it now appears to have found a workaround.
Facing EU regulators, Meta also announced that the next multimodal version of Llama would not be launched in Europe, though the text-based version remains accessible. The same applies to its smart glasses: multimodal features, which allow users to request information about what they see, remain unavailable on the continent.
