Meta to use AI chat data for ads without user opt-out
Meta has announced that it will use users' conversations with its AI assistant to personalize advertisements on Facebook and Instagram. The company will not provide an opt-out option for this form of ad targeting. This policy begins in the European Union and will expand to other regions.
Meta, the parent company of Facebook and Instagram, revealed in October 2025 that interactions with its AI chatbot, Meta AI, will inform targeted advertising across its platforms. Powered by the Llama 3.1 model, Meta AI is integrated into apps like Messenger, WhatsApp, and the main Facebook and Instagram feeds, allowing users to engage in natural conversations for tasks such as planning or information queries.
According to Meta's updated privacy policy, 'We’re using your interactions with Meta AI to improve your experience on our platforms, including by showing you more relevant ads.' This means data from these chats—such as topics discussed or preferences expressed—will be analyzed to tailor ads, similar to how other user behaviors like likes or searches are already utilized. However, unlike some ad preferences where users can adjust settings, there is no specific opt-out for AI-derived targeting. Users can delete individual chats or all conversation history, but this does not prevent the data from being used for ad personalization once processed.
The rollout starts in the EU, where stricter data protection laws like the GDPR apply, and will gradually extend globally. Meta stated in its announcement: 'Interactions with Meta AI may be used to personalize ads.' This move aligns with the company's long-standing reliance on user data for revenue, which generated over $130 billion from advertising in 2024 alone. Privacy advocates have raised concerns, noting that AI chats often involve sensitive personal details, but Meta maintains that the practice enhances user relevance without new data collection beyond existing permissions.
Background on Meta AI shows it launched in 2023 as a competitor to tools like ChatGPT, emphasizing seamless integration into social experiences. While the policy clarifies ad usage, it does not alter how Meta AI handles data for model training, which already excludes personal identifiers. No timeline for full global implementation was specified, but sources indicate it could occur within months following the EU debut.