WhatsApp’s new AI won’t read your messages—Meta says

Written by

Published 1 May 2025

Fact checked by

We maintain a strict editorial policy dedicated to factual accuracy, relevance, and impartiality. Our content is written and edited by top industry professionals with first-hand experience. The content undergoes thorough review by experienced editors to guarantee and adherence to the highest standards of reporting and publishing.

Disclosure

icon

Meta announced it is adding artificial intelligence (AI) to WhatsApp messages without breaking the app’s encryption. The new feature called “Private Processing” will let WhatsApp users access AI features like message summarization while keeping their messages private.

The feature uses specially protected cloud servers that process messages without allowing Meta, WhatsApp, or others to see the content. It works similarly to Apple’s recent Private Cloud Compute for Apple Intelligence.

    “We think building a private way to use AI is important, because people shouldn’t have to switch to a less-private platform to have the functionality they need,” WhatsApp head Will Cathcart told WIRED.

    Private Processing arrives as Meta competes with Apple and other companies adding AI to messaging. The company must balance adding new features while protecting WhatsApp’s reputation for privacy and security.

    Users will need to turn on the feature themselves. It will be off by default when it launches in the coming weeks. For extra safety, a separate setting called “Advanced Chat Privacy” can block AI from working in specific conversations.

    Security expert Chris Rohlf, who helped develop the system, emphasized user choice: “This wasn’t just about managing the expansion of that threat model… it was about careful consideration of the user experience and making this opt-in.”

    The system uses several safety measures. These include third-party relays to hide user locations, temporary encryption keys, and automatic deletion of all data after each session. Meta has released a detailed summary blog of the technology.

    Johns Hopkins cryptographer Matt Green reviewed the system and expressed mixed feelings. “I believe WhatsApp when they say that they’ve designed this to be as secure as possible, and I believe them when they say that they can’t read your texts. But I also think there are risks here.”

    Green pointed out that “more private data will go off device, and the machines that process this data will be a target for hackers and nation state adversaries.”

    To build trust in the system, Meta plans to publish parts of the system’s code and invite security researchers to test it through its Bug Bounty Program.

    Meta plans to use Private Processing for future AI features in WhatsApp and possibly other apps. The announcement follows other recent WhatsApp privacy upgrades, including the Advanced Chat Privacy feature, which gives users more control over how their messages are used.