Meta’s new Incognito Chat arrived this week with a bold promise: ask its AI anything, health questions, money problems, relationship troubles, and even meta’s own engineers cannot read what you typed.
CEO Mark Zuckerberg announced the feature for WhatsApp and the Meta AI app. In Nigeria, where WhatsApp is not just a messaging app but the infrastructure for mobile banking alerts, business negotiations, and money transfers, the question is not just whether Incognito Chat works. It is if Meta has earned the right to ask for this kind of trust.
READ: OpenAI's Daybreak Joins the AI Cybersecurity Fight
What Meta Incognito Chat Actually Does
Meta built Incognito Chat on its Private Processing framework, the cloud infrastructure WhatsApp has used since last year to route AI requests through isolated secure environments. When a user activates the mode, their prompts travel into Trusted Execution Environments that Meta says neither its staff nor outside parties can access.
Nothing is stored. Close the session, and the conversation is gone. Meta says activation is always a deliberate user choice; nothing reroutes silently into the private system.
At launch, Incognito Chat handles text only. No images, voice notes, or files. Meta’s safety systems still scan for illegal or harmful content in the background, even inside the encrypted environment.
Meta is also building Side Chat, a separate mode that lets someone ask the AI about a conversation in the background, summarise it, or decode it without the other person seeing anything. Side Chat runs on the same private processing infrastructure.
Why Meta’s Timing Is a Trust Problem
The announcement landed at the worst possible moment.
Days earlier, Meta ended end-to-end encryption for Instagram direct messages globally, effective May 8. Users woke up to notifications that their encrypted DM history would be deleted unless they saved it manually. Meta’s explanation: very few people had opted into the encrypted DM feature in the first place.
Privacy advocates called out the contradiction immediately. Meta stripped encryption from one platform. Days later, it unveiled an encryption-forward AI product on another.
In Nigeria, that contradiction lands harder. WhatsApp became the trusted messaging channel precisely because it offered encryption that local SMS and telecom channels do not. Moving AI conversations into that same space while Meta weakens encryption elsewhere is not a reassuring context. Users who use WhatsApp for bank alerts and business negotiations are not casual users. They have something to lose.
READ: AI Ethics, Religion, and Why You Can't Leave It to Silicon Valley
What Security Experts Say Meta Cannot Guarantee
Trusted Execution Environments are a real and meaningful security tool. But no cloud system carries a permanent guarantee.
Cybersecurity researcher Alan Woodward, speaking to the BBC, raised a concern that Meta’s pitch does not address. If conversations are inaccessible even to Meta, outside researchers also lose the ability to audit harmful outputs. An AI that gives dangerous medical advice or facilitates financial fraud inside a private session creates accountability gaps that regulators, including Nigeria’s Federal Competition and Consumer Protection Commission, will eventually have to confront.
Governments want to see inside these systems. Users want the opposite. Meta is running toward the users and hoping regulators do not catch up.
ChatGPT, Google Gemini, and Claude all retain some conversation data for safety checks, abuse prevention, or model improvement. Meta claims to keep none. That claim puts it ahead of rivals on this specific point, though no outsider can verify it independently.
READ: How to Remove Private Number on Android: Every Method That Works
WhatsApp Is Carrying the Weight of Meta’s AI Ambitions
WhatsApp has over three billion users globally. It runs on the Signal Protocol and built its encryption reputation before Meta started pushing AI features into it. That reputation is real, and it is what Meta is now trading on.
WhatsApp’s existing goodwill either extends to Incognito Chat or gets spent by it. That is the question Meta’s history makes genuinely difficult to answer.
The technical architecture is credible. The privacy framing is coherent. But a company that ended encrypted Instagram DMs one week and launched encrypted AI chat the next is not a company that has earned automatic trust. Nigerian users in particular, who have built real financial and professional lives inside WhatsApp, deserve more than a promise.
Meta insists Incognito Chat changes the equation. Users will decide whether the company that changed the Instagram equation gets to change this one, too.
Sign up for the GizPulse newsletter — free, fast, and always worth your time.
