Signal blocks screenshots on mobile to keep AI agents from reading chats
The messaging app is restricting screenshots on Android and iOS, saying built-in AI assistants undermine its privacy guarantees, and it is urging OS makers to give developers better controls.
· 2 min read
- Signal is blocking screenshots in its Android and iOS apps to keep AI assistants from reading on-screen content.
- President Meredith Whittaker calls the fix incomplete and says operating system makers should give developers simple tools to limit AI data access.
- Users can turn screenshots back on in settings, but Signal warns this may expose their chats and their contacts' messages to AI agents.
Signal is restricting screenshots in its mobile apps on Android and iOS, a move meant to stop AI agents built into the operating system from reading users' private messages, according to a report from The Verge published on 20 November 2026 at 12:30 UTC.
Signal president Meredith Whittaker explained the decision in a blog post. She said the agentic AI features now built into operating systems, which analyze what appears on a user's screen to learn about them, are undermining Signal's ability to guarantee privacy on its own. She also said such agents are hostile to privacy.
The block applies to Android and to iOS 27. It is intended to prevent Siri, Gemini and other Android AI assistants from reading on-screen content in the app. Users who want screenshots back can re-enable them in Signal's settings. Whittaker warned that doing so could expose both their own Signal content and that of their contacts to these AI agents, and would break the app's privacy guarantees.
Whittaker acknowledged that the measure is incomplete. Blocking screenshots does not cover other routes an AI agent might take to reach app data, such as pulling it from the device's filesystem. Signal also cannot apply the same protection on macOS at present.
According to Whittaker, the underlying problem is that Signal and other app developers lack strong tools to stop AI agents from undermining privacy at the application level. She said that even when Signal can detect and block AI access to data, Google, Apple, Microsoft and others offer no easy way to restrict it without drastic compromises to the core user experience.
The report notes that Signal took a similar step last year to keep Microsoft's Recall feature from capturing secured chats. The service says it is struggling to build a better fix because little information is available about how AI agents actually reach application data. It is asking Apple, Google and Microsoft for simple technical tools and accurate documentation for developers.
Signal also wants more protection against third-party agents, citing Meta's Muse as an example, which can use deep accessibility settings in apps as a kind of skeleton key to reach application data.
Whittaker described the current integration of AI agents into operating systems, without such remedies, as an existential threat to privacy at the application layer. She said Signal is the canary in the coal mine, but that the issue affects the whole ecosystem.