WhatsApp’s Permanent Meta AI Shortcut Raises Privacy and Control Concerns

Meta has begun rolling out a permanent Meta AI shortcut inside WhatsApp, represented by a small blue‑pink‑green circle in the lower‑right corner of the Chats screen. Tapping it opens a chatbot powered by Meta’s Llama 4 language model that can answer questions, provide weather updates, suggest ideas and generate web links. Although Meta repeatedly calls the feature “optional,” users cannot remove the icon or disable the function, prompting a wave of complaints on X, Bluesky and Reddit and criticism from commentators such as Guardian columnist Polly Hudson.

A WhatsApp spokesperson told the BBC the firm regards the shortcut as no different from fixed interface elements such as Channels or Status and added that it is “always listening to feedback.” The tool is being released gradually and is not yet visible to all users; Meta cautions that availability varies even within the same country. Before first use, a disclaimer explains that only messages sent to the chatbot are visible to Meta, while ordinary person‑to‑person chats remain end‑to‑end encrypted.

Privacy specialists are unconvinced. Dr Kris Shrishak argues that Meta is leveraging its dominant position to extract fresh conversational data after the legality of training Llama on web‑scraped content—and allegedly millions of pirated books—was challenged in court. Several authors have already sued Meta over the practice. The UK Information Commissioner’s Office says it is monitoring how Meta AI handles personal data, especially that of minors. Critics contend that users should never be compelled to interact with an AI system and warn that any information shared with Meta AI could be retained or reused. The controversy follows Microsoft’s recent reversal on its own always‑on Recall feature, highlighting growing consumer resistance to un‑removable AI assistants embedded in everyday software.