WhatsApp has launched a new AI feature that Meta claims is completely optional. However, users report they cannot disable it. In some countries, a colorful circle with the “Meta AI” symbol now appears in the chat view. Tapping it opens a chatbot that answers questions, provides ideas, and shares information. Many users find the feature intrusive, especially since there is no way to turn it off. Privacy experts warn that the supposedly optional tool raises serious concerns about handling personal data.
What is Meta AI and how does it work?
Meta AI is a chatbot powered by artificial intelligence, currently being introduced in selected regions on WhatsApp. It runs on Meta’s Llama 4 language model and is designed to provide quick answers, support creative thinking, and help with learning. The feature is also integrated into Facebook Messenger and Instagram, which are both owned by Meta.
Users who already have access see a search bar at the top of the chat window that says “Ask Meta AI or search.” By tapping the round, colorful icon at the bottom right corner of the screen, a chat window opens where users can ask questions like “What’s the weather?” and receive instant answers with additional links.
However, problems have been reported. In one test, a user asked about the weather in Glasgow. The AI responded with information, but the links led to weather data for Charing Cross in London.
Users complain about lack of deactivation option
Although Meta says the feature is voluntary, many users say they cannot remove or turn it off. On platforms like X (formerly Twitter), Bluesky, and Reddit, users have expressed frustration. Many feel forced to keep the AI feature even if they never use it.
British columnist Polly Hudson criticized that there is no option to disable the chatbot. AI and privacy experts are also concerned. Dr. Kris Shrishak, a leading expert on artificial intelligence and data protection, stated that Meta is using its dominant market position to turn people into test subjects, leaving them without a real choice.
Dispute over training data
Another major criticism is about the data used to train Meta AI. According to Dr. Shrishak, much of the training data came from publicly available websites and possibly even illegal sources. The controversial site Library Genesis, which offers access to millions of copyrighted books, may have been used in the process.
Writers’ organizations across the world are demanding action. Many authors have already filed lawsuits against Meta. They believe their copyrighted work was used without permission. Meta has not yet responded to these claims.
Data protection authorities to review the feature
Meta claims that Meta AI only accesses information that users share directly with the bot. Private chats, the company says, remain end-to-end encrypted and are not affected.
Despite these claims, a national data protection authority has announced it will monitor Meta AI’s usage closely. Officials emphasized that AI tools must follow existing data protection laws, especially when handling sensitive content or information from minors.
Dr. Shrishak explained that when a person chats with Meta AI, Meta becomes one of the two participants in the conversation. This makes it very different from private messages between individuals.
Meta advises users not to send confidential information or personal data of others to the AI.
Controversial rollout despite being labeled optional
Although Meta AI is promoted as a voluntary tool, many users feel it is being forced upon them. The lack of a deactivation option and unanswered questions about privacy have led to widespread dissatisfaction. While Meta continues to highlight the tool’s usefulness and safety, many users have lost trust.