Signal Calls OS-Level AI an 'Existential Threat' and Blocks Screenshots to Fight It

Signal Calls OS-Level AI an 'Existential Threat' and Blocks Screenshots to Fight It

Summary

Signal said it will block agentic AI features in iOS 27 and automatically enable Android’s “screen security” setting when an assistant tries to read the app’s screen. The protections aim to stop AI systems from capturing Signal conversations, which Signal says would put messages outside its encryption guarantees. Users can disable the setting, but Signal warns this may expose messages. The company said the fix does not prevent access through methods such as a device’s filesystem and cannot yet be implemented on macOS. Signal urged Apple, Google and Microsoft to provide clear ways for apps to opt out of AI data access. It also warned about third-party agents using accessibility permissions, citing a report that Meta’s Muse accessed a journalist’s private iMessages.

Latest News!