WhatsApp Just Added an ‘Incognito’ Mode for AI Chats – Here’s How to Use It
In mid-May 2026, Meta began rolling out a new privacy control for WhatsApp users: an “incognito” mode specifically for conversations with the app’s built-in AI chatbot. The move comes as more people interact with generative AI inside messaging apps and as regulators and users alike ask tougher questions about how chat data is used.
If you use WhatsApp’s AI features – or have been hesitant to try them because of privacy worries – this change is worth understanding. Here is what the incognito mode does, how to turn it on, and what limits remain.
What Happened
Meta announced the feature on May 13, 2026, after weeks of testing. According to reports from ABC News and Yahoo Tech, the incognito mode prevents WhatsApp from using the content of your AI conversations to train or improve Meta’s AI models. When the mode is active, those chats are still stored on your device and can be seen by you and Meta (as part of normal WhatsApp operation), but they are excluded from the company’s model‑training pipeline.
The rollout covers both Android and iOS users globally. As with many WhatsApp updates, the feature may take a few days to reach everyone, depending on your region and app version.
Why It Matters
AI chatbots in messaging apps raise a fundamental privacy question: who gets to see what you type? By default, WhatsApp’s AI assistant – like most consumer AI tools – can use your conversation history to improve its responses and to train future models. For many people, that is an acceptable trade‑off; for others, it is a deal‑breaker.
The incognito mode gives users a straightforward way to opt out of that data‑use agreement without abandoning the AI feature entirely. It is not end‑to‑end encryption for AI chats (WhatsApp’s regular messages remain end‑to‑end encrypted, but the AI component works differently). Rather, it is a training‑data exclusion mechanism.
Still, the feature addresses a specific and growing concern. In a 2025 Pew survey, roughly half of U.S. adults said they had avoided using a generative AI tool because of privacy or data‑sharing worries. WhatsApp’s incognito mode does not solve all of those worries, but it is a step toward giving users more control.
What Readers Can Do
How to enable incognito mode for AI chats
- Open WhatsApp and tap the three‑dot menu (Android) or the Settings tab (iOS).
- Go to Settings > Privacy.
- Look for a section labelled AI Chats or AI Privacy. (The exact name may vary by version.)
- Toggle on Incognito Mode (or a similarly labelled switch).
- A confirmation message will appear explaining that your AI chats will no longer be used for training. Tap Confirm.
Once enabled, the setting applies to all future AI conversations within the app. Chats from before you turned it on may still have been used for training, so if you have existing AI conversations you want to protect, consider deleting them after enabling the mode.
What incognito mode does not do
- It does not change how AI chats are encrypted. They are still sent to Meta’s servers for processing. End‑to‑end encryption applies only to person‑to‑person messages, not to AI interactions.
- It does not stop Meta from collecting metadata such as your phone number, IP address, device info, or how often you use the AI feature. That data is still subject to Meta’s standard privacy policy.
- It does not prevent Meta from storing your AI chats on its servers. The chats remain visible to Meta but are simply excluded from model‑training datasets.
Comparison with other platforms
- Signal does not have a built‑in AI chatbot, so this type of privacy control is not applicable.
- Telegram allows third‑party AI bots, but there is no equivalent “incognito” setting; users must rely on each bot’s individual privacy policy.
- iMessage integrates with Apple Intelligence (or Siri), and Apple states that on‑device processing limits data exposure. However, users cannot easily opt out of model training for queries that do reach Apple servers.
WhatsApp’s incognito mode is one of the more explicit user‑facing controls for AI training data in a mainstream messaging app today.
Sources
- ABC News – “Meta launches WhatsApp ‘incognito’ mode to address privacy concerns for AI chats” (May 13, 2026)
- Yahoo Tech – “Meta launches incognito AI chat mode on WhatsApp” (May 13, 2026)
- Pew Research Center – “Which U.S. adults have used generative AI?” (2025)
Note: Exact menu labels may change as Meta refines the feature. Check WhatsApp’s official help centre for the latest instructions if the steps above do not match your app version.