WhatsApp Launches Private AI Chat Mode With No Server Logs
BBC Business reported Wednesday that WhatsApp has introduced a WhatsApp incognito AI chat feature that prevents even Meta itself from reading user conversations. The new mode stores no server-side logs and wipes past exchanges from the user’s own view once a session ends.
What the Feature Actually Does
WhatsApp head Will Cathcart said many users felt uneasy sharing sensitive topics with an AI they knew was being monitored. Health questions, relationship concerns, and financial matters were cited as areas where users sought more discretion. Cathcart described the underlying protection as functionally equivalent to the platform’s existing message encryption, though built on a different technical architecture. Meta CEO Mark Zuckerberg called it the first major AI product to eliminate conversation logs from company servers entirely.
Background: Meta AI’s Rocky Reception
When Meta first embedded its AI assistant into WhatsApp last year, a vocal segment of users pushed back. The chief complaint was that the feature could not be switched off. Despite that friction, Zuckerberg announced in May 2025 that Meta AI had crossed one billion active users across its family of apps. Meta has blocked rival AI chatbots from accessing WhatsApp’s infrastructure, meaning Meta AI is the only assistant available to the platform’s global user base.
Also Read: Meta Shares Slide as $145 Billion AI Spend Rattles Investors
Accountability Gap Worries Experts
Professor Alan Woodward, a cyber security specialist at Surrey University, told the BBC the additional system introduced minimal new risk to WhatsApp’s core security. However, he flagged a more fundamental concern. If a conversation contributes to harm, self-harm, or death, no retrievable record would exist on either side. That evidentiary gap could shield Meta from scrutiny and complicate any legal proceedings. Several AI companies, including OpenAI and Google, already face wrongful-death lawsuits tied to chatbot interactions. Woodward acknowledged the genuine privacy case for protecting sensitive AI queries but said users would be placing extraordinary trust in a system with no external check.
Investor Pressure Looms Over Every Feature Decision
Analyst Susannah Streeter of investment platform Wealth Club noted that Meta is on course to spend roughly $145 billion on AI infrastructure this year. Shareholders have grown nervous about the pace of that outlay. Each new product release is therefore being weighed against whether it can meaningfully deepen engagement and reinforce Meta’s advertising and commerce dominance. Incognito mode, in that context, is as much a retention tool as a privacy one. The feature initially supports text input only, with image processing excluded while guardrails are calibrated.
Read Next: Meta Investigated Over AI Having Sensual Chats With Children
