A feature called Incognito Chat is coming to Meta AI and WhatsApp soon, according to Mark Zuckerberg.
Meta’s boss announced the feature on his Facebook page and described it as allowing users a “completely private way” to interact with the company’s AI assistant.
Meta made $14 million just off of these scam ads
“This is the first major AI product where no log of your conversations is stored on servers,” Zuckerberg wrote.
He said this feature is like end-to-end encryption, which means “no one can read your conversations, not even Meta or WhatsApp.”
Although the conversations cannot be read by the platform itself, they also disappear when the user expires.
Mashable 101 Favorites: Vote for your favorite creator today!
“To get the most out of personal intelligence, we’ll all need ways to discuss sensitive topics in ways no one else can,” Zuckerberg wrote.
Incognito Chat privacy and security
The disappearing chats raise security questions that weren’t addressed in Zuckerberg’s blog and Meta in Incognito Chat.
While complete privacy may prevent users from asking sensitive questions about their health, finances, or work, it will also protect Meta from knowing where users may need emergency help or intervention.
Mashable Light Speed
For example, conversations with Meta AI on WhatsApp that indicate a user may be thinking about harming or killing themselves may trigger a person’s review, according to Mashable’s analysis. The same is true of violent discussions.
These messages could not be identified through Incognito Chat, nor would there be any record of them.
Meta said it uses safeguards designed to reject potentially harmful information, and that Meta AI will not match malicious requests. Additionally, users who repeatedly send malicious commands will be temporarily banned, according to the company.
Both of these situations – suicidal behavior and acts of violence in society – are the subject of lawsuits and criminal inquiries against major AI companies.
OpenAI has been sued many times by bereaved families of users. They suspect that OpenAI’s ChatGPT trained their loved one to kill themselves. OpenAI denied the allegations in another case involving a 16-year-old child.
Separately, the attorney general of the state of Florida recently opened a criminal investigation into whether ChatGPT provided “significant” tips to the gunman who allegedly killed two people and five others in an April 2025 shooting.
Google, maker of chatbot Gemini, was sued for wrongful death earlier this year by an elderly man’s family after it was alleged that Gemini convinced him to kill himself.
“Our models generally perform well in these types of challenging conversations and we spend significant resources on this, but unfortunately AI models are not perfect,” Google said in a statement following the allegations.
The lawsuits against Google and OpenAI draw heavily on user chat logs.
Can teenagers use Incognito Chat with Meta AI?
Meanwhile, in an effort to strengthen protections for new users of Meta AI, the company recently released a feature that allows parents to view their chat topics with the AI.
Incognito Chat is intended for users 18 and older, according to Meta. Users will be prompted to verify their age before using this feature. If required by law, Meta will use additional age verification methods to verify that a user is an adult.
Sarah Gardner, CEO of the Heat Initiative, an advocacy group focused on internet safety and corporate accountability, expressed concern about Incognito Chat, especially given Meta’s previous release of AI chatbots that allow for “sensational” conversations with children.
“The new features announced today should raise alarm bells for parents,” Gardner said in a statement to Mashable. “We’re not optimistic about Meta’s record of age verification, so they need to answer a lot of questions about how they’re going to ensure the safety of children.”
