Signal announced on Friday that it is limiting the ability of users to take screenshots within its secure messaging application. This move targets operating systems that integrate agentic AI, which President Meredith Whittaker described as an existential threat to the platform.
Whittaker explained that modern operating systems for Windows, Android, and macOS or iOS now allow AI agents to scan information displayed on users’ screens. This capability undermines Signal’s ability to independently guarantee privacy for its users.

Operating System AI Risks
According to Whittaker, these AI capabilities capture sensitive data, store it, and make it accessible to local databases or remote servers used by the operating system’s AI agent. Once the data is captured by the OS, it falls outside of Signal’s control, imperiling the privacy of all participants in a conversation.
Signal had previously addressed this issue by blocking screenshots of its Windows application by default. The new update extends similar restrictions to Android and iOS platforms.

Privacy Guarantees and Limitations
Users retain the option to disable this setting if they choose. However, Whittaker warned that doing so would break the app’s privacy guarantees. She also acknowledged that this measure is an incomplete solution, as it does not address other potentially sneaky ways AI agents could access application data.

The restriction aims to close a specific vulnerability where screen content is intercepted by system-level AI tools. This step highlights the growing tension between advanced operating system features and end-to-end encryption principles.
Source: TechCrunch

