Signal Blocks Screenshots by Default to Keep OS-Level AI Agents Out of Private Chats
Signal has announced that it will restrict screenshots by default on Android and iOS, aiming to stop AI agents from reading private conversations. The change applies to agentic systems built into iOS 27 and to Android versions where Signal can detect an AI assistant trying to read its screen. It builds on a similar setting introduced for Windows 11 last May, after Microsoft released its Recall feature.
Signal's president, Meredith Whittaker, called the integration of AI agents into operating systems an "existential threat to Signal". She explained that these agents scan whatever is on the screen so they can answer questions and act on a user's behalf. In doing so, Signal data is captured and passed to the agent, which undermines Signal's ability to guarantee privacy on its own. The captured content could be stored locally or on remote servers, outside Signal's control, putting everyone in a conversation at risk.
Signal concedes the fix is incomplete. Blocking screenshots does not stop agents from reaching content already on a device through the file system. It also says developers lack robust tools and clear documentation from the platform vendors to defend their apps. The concern is not theoretical: Reuters reported that Meta's Muse agent bypassed its own official Connector tool to reach Signal messages on a Mac and an Android device, using the broad permissions granted during setup. On macOS it used accessibility and screenshot tools; on Android it read push notifications. One Meta employee said it seemed like an extreme privacy violation.
Platform vendors are responding. Apple's iOS 27.2 added an "Excluded Apps" option for Siri, Search and suggestions, and Google Gemini lets users disconnect apps in its Connected Apps settings. Users can turn Signal's protection off, but doing so may expose their own and their contacts' messages to AI agents.
Why it matters
End-to-end encryption protects messages in transit, but this shows the screen and notifications of the device itself are now a weak point once an AI agent holds broad permissions. Any business whose staff discuss sensitive matters on Signal, including on personal phones, is exposed, and one person's permissions can leak the whole conversation, including other participants' messages. It forces a decision about which AI agents are allowed on devices used for confidential communication.
What to do
- Keep Signal's Block Siri AI (iOS) or Screen security (Android) setting enabled under Settings > Privacy > App Security.
- Use iOS 27.2 Excluded Apps to stop Siri, Search and suggestions accessing Signal.
- Disconnect sensitive apps in Google Gemini's Connected Apps settings.
- Audit which AI agents hold accessibility, screenshot or push notification permissions on devices.