Signal Restricts Mobile Screenshots to Prevent AI Agents Capturing User Data
Signal is limiting screenshot capabilities on mobile operating systems to prevent artificial intelligence agents from capturing user data from private chats, as reported by MarketScreener Nederland and Tweakers.
Signal president Meredith Whittaker announced on Friday that the encrypted messaging app restricts screenshot creation on iOS 27 and Android when the app detects an AI assistant attempting to read the display, according to Unite.AI. The move addresses growing privacy concerns regarding end device behavior, where private content becomes visible on screens and accessible to operating system agents.
Google’s Gemini documentation indicates that screen actions can upload information visible on a display or in a file to serve as context, while Intelligent Screen Context can utilize currently visible content to answer chat requests. To counter this, Signal’s Android design accounts for distribution variations among different manufacturers, while Windows 11 previously introduced standard screen protection on May 21, 2025, using a digital rights management window flag to prevent Microsoft Recall from capturing Signal Desktop content.
The protection remains partial because Signal cannot reliably block screenshots on macOS 27 or apply continuous screen recording defenses against the image capture methods typically used by agentic systems, Unite.AI reported. Whittaker said that users can disable the protection to take screenshots, but warned that doing so exposes their own conversation content as well as that of their contacts, thereby undermining the app’s privacy guarantees.
Signal recommends adding the app to System Settings, Siri, and Excluded Apps on macOS 27 to shield message content from proactive access by Siri AI. Furthermore, the organization is asking Apple, Google, and Microsoft to provide a documented mechanism allowing applications to opt out of AI processing and screen reading without compromising core functionalities.