WhatsApp推Meta AI 無痕對話：號稱連Meta都看不了，但有一大隱憂

WhatsApp has launched “Meta AI Incognito Chat,” ensuring conversations are not stored and are inaccessible to officials. Experts are concerned this could make AI accountability difficult. Meta says that at the initial stage it supports only text and uses conservative protections.

Meta AI Incognito Chat Debuts, Created in Response to Privacy Demands

As generative AI chatbots rapidly become widespread, talking with AI has gradually become part of everyday life. However, many questions involve extremely sensitive content—such as providing private financial, personal, health, or work information—at which point privacy needs also arise.

For this purpose, WhatsApp has launched “Meta AI Incognito Chat,” claiming it is a brand-new feature for fully private conversations with AI.

Based on private-processing technology, Meta AI Incognito Chat enables users to have private chats with Meta AI without others being able to see them. User messages are processed in a secure environment, and even Meta itself cannot access them.

These conversations are not stored, and the system is set by default to automatically make messages disappear, giving users a space to think freely and explore ideas without anyone snooping.

Image source: WhatsApp WhatsApp adds the Meta AI Incognito Chat feature to meet people’s privacy needs

Six Core Technologies of Meta AI Incognito Chat

According to Meta’s technical white paper, Meta AI Incognito Chat mainly combines the following six core technologies to ensure the privacy and security of conversation data:

1. Confidential Computing Hardware: Uses AMD CPUs and Nvidia GPUs that support confidential computing to build a Trusted Execution Environment (TEE), ensuring the computing tasks are isolated at the hardware level, including Meta or the host operating system, all of which cannot access the data being processed.
2. Authenticated and Encrypted Communication: The system uses RA-TLS to provide end-to-end encryption, ensuring that only the user device and the private-processing nodes can decrypt, and it also uses hardware authentication to verify whether the server is running an unmodified software stack.
3. Artifact Transparency: To prevent malicious software deployment, the system publishes important components such as binary files and model weights to third-party public transparency logs, allowing clients and researchers to verify the authenticity of the code being executed.
4. Secure Software: Inside the TEE, a multi-layered protection software stack is used, and applications are containerized to limit the attack surface and strictly control data-leakage pathways.
5. Anonymous Routing: The system uses an anonymous credential service and a stealth HTTP technology. It performs routing through third-party relay proxy servers, hiding users’ IP addresses from Meta and de-identifying users to prevent attackers from targeting specific users’ data.
6. Short-lived and Stateless Data Processing: The coordinator and predictor inside the system are designed to be stateless. After processing a request and returning the result, it discards the conversation data in memory to ensure there is no access to retained historical records.

External Concerns: What If Meta AI Is Involved in Unnatural Deaths?

According to a report by the BBC, most AI companies currently store chatbot usage data and use it to train future product models. WhatsApp executive Will Cathcart explained that the technology and platform behind WhatsApp’s Incognito Mode differ from the end-to-end encryption used to protect other messages, but both are equally effective.

Surrey University cybersecurity expert Alan Woodward also said that introducing a second system poses very low risk to WhatsApp’s existing security.

However, external concerns remain that Incognito Mode could conceal AI failures or abuse. For example, multiple AI companies, including OpenAI and Google, have previously faced lawsuits related to unnatural deaths.

Woodward believes this could lead to a lack of an accountability mechanism for AI responses, because automatically disappearing messages cannot be retrieved by users or Meta. That means if someone’s conversation results in harm or death, it may be impossible to find relevant evidence.

Cathcart responded that the initial version of Meta AI Incognito Chat will handle only text and will not support images. At the same time, Meta AI’s safety protection mechanisms will be conservative and will refuse to answer requests that could be interpreted as harmful or illegal.

In addition, WhatsApp has currently blocked other AI chatbots from accessing its system. Therefore, among the platform’s billions of users, the only AI they can interact with is Meta’s own product.

Related Reading:

Train AI with Employees! Meta Launches Internal Tracking Tools, Recording Employees’ Mouse Clicks and Keystrokes Meta Doubles Down on AI: Zuckerberg Uses Claude to Write Code, Employees Kick Off a Token Consumption Battle to Push KPIs