Signal Pulls the Blinds: No More Easy Screenshots While AI Agents Are Watching
Signal, the encrypted messaging app beloved by journalists, activists and privacy fans, is now restricting screenshots inside its Android and iOS apps. The target isn't your nosy friends. It's the AI agents that Apple, Google and Microsoft are building straight into their operating systems.
Signal president Meredith Whittaker laid out the problem in a blog post on Friday. Many new AI assistants work by reading what's on your screen. So a message sent with perfect end-to-end encryption can get captured the moment it's displayed, then stored or shipped off to whatever database or server the OS-level agent uses. Once that happens, Signal can't protect it, and neither can anyone else in the chat. Whittaker called it an existential threat to the app's mission.
Signal is extending a defense it first used on Windows, where it blocked screenshots by default to keep Microsoft's Recall feature from snapping chats. According to reports on the rollout, iPhones running iOS 27 get screenshot blocking to keep Siri's AI out, while on Android the app's Screen Security kicks in when it detects an AI assistant trying to read the screen. Signal also called out third-party agents like Meta's Muse that lean on accessibility permissions.
There are catches. Users can switch the protection off, though Whittaker warned that doing so breaks Signal's privacy guarantees. She also admitted the fix is incomplete, since AI agents may have sneakier ways into app data. Her bigger ask is for Apple, Google and Microsoft to give developers simple tools to keep AI agents away from sensitive apps.
Photo Credits: techspot.com
