Signal Blocks Screenshots on iOS 27, Restricts AI Screen Capture on Android
Signal announced Friday it will block screenshots on iOS 27 and restrict AI screen capture on Android, moving to shield encrypted chats from operating-system assistants that can read decrypted content off the display.
The Tech TL;DR:
- Signal blocks screenshots by default on iPhones running iOS 27 via a new “Block Siri AI” toggle; on Android, “Screen security” activates automatically when an AI assistant attempts screen access.
- The protections address a gap where end-to-end encryption stops at the screen: OS-level agents like Siri AI, Gemini, and Meta’s Muse can ingest decrypted messages from notifications, accessibility APIs, or screenshot streams.
- Signal acknowledges the defense is partial — macOS screenshots cannot be reliably blocked, and filesystem or notification access remains a vector — and calls on Apple, Google, and Microsoft for documented opt-out APIs.
Encryption ends at the display
End-to-end encryption secures messages in transit and at rest, but the recipient’s device must eventually render plaintext. Once rendered, that text exists in the UI layer — chat bubbles, notification previews, accessibility trees — where OS-integrated AI agents can capture it.
Signal President Meredith Whittaker wrote that agentic AI integrations are “changing this environment in a way that’s making it increasingly hostile to privacy” and pose “an existential threat to Signal, and our ability to responsibly uphold our mission,” according to the organization’s blog post.
Reuters, working with Signal and the AI Now Institute, documented Meta’s Muse assistant reading a Signal message on Android via notification access and on macOS via accessibility tools and screenshots, Mashable reported. The Meta employee who received the Muse reminder said she “definitely did not grant it access to other apps, or recall being asked about this.” Meta told Reuters Muse accesses other apps only through permissions users explicitly grant.
iOS 27: Screenshot blocking by default
On iOS 27, Signal introduces a “Block Siri AI” setting enabled by default. The mechanism uses Apple's screenshot prevention APIs — similar to those used by DRM-protected apps — to return blank frames when the system or an assistant attempts a capture.
Unite.ai noted Apple said on Sept. 14, 2026 that Siri AI would begin rolling out in English beta with 2027 software releases, including onscreen awareness to answer questions or take actions related to screen content.

Users can disable the block in Signal settings, but the app warns doing so “will potentially expose your (and your contacts’) Signal content to these AI agents, and break our privacy guarantees,” per the Signal blog.
Android: Detection-triggered screen security
Android’s fragmented vendor distributions prevent a uniform default block. Instead, Signal enables “Screen security” — a WindowManager flag (FLAG_SECURE) — when it detects an AI assistant attempting to read the screen. Google’s Gemini documentation states screen actions can upload visible information for context, while Intelligent Screen Context uses currently visible content to answer chat requests, Unite.ai reported.
The detection heuristic relies on identifying known assistant packages and accessibility service signatures. Vendors that customize the assistant stack may evade detection, leaving a protection gap Signal cannot close without platform-level APIs.

Signal blocks screenshots on Windows, macOS still vulnerable
Signal cannot reliably block screenshots on macOS 27. The organization recommends users add Signal under System Settings > Siri > Excluded Apps to limit proactive access by Siri AI, Unite.ai reported. On Windows 11, Signal enabled Screen security by default on May 21, 2025, using a digital rights management window flag so screenshots appear blank and Microsoft Recall cannot capture contents, according to Signal Messenger blog and Unite.ai.
Confidential computing approaches — trusted execution environments, secure enclaves — do not address these core concerns because the data is captured after decryption by the OS compositor, Whittaker wrote.
Platform asks: Documented opt-out APIs
Signal is asking Apple, Google, and Microsoft for a documented way to opt apps out of AI access across screen reading, filesystem access, notifications, APIs, and other OS frameworks “without undermining core utility,” Unite.ai reported.

Meta maintains Muse operates within granted permissions. Signal counters that broad accessibility permissions — designed for disability access — function as “a skeleton key that provides access to AI agents,” Whittaker wrote.
Implementation reference: FLAG_SECURE on Android
// In Activity.onCreate() before setContentView()
getWindow().setFlags(
WindowManager.LayoutParams.FLAG_SECURE,
WindowManager.LayoutParams.FLAG_SECURE
);
// Prevents screenshots, screen recording, and non-secure display projection
Signal toggles this flag dynamically when its detector identifies an active AI assistant service. The iOS equivalent uses UITextField or UIView secure text entry properties combined with screenshot detection callbacks introduced in iOS 27 SDK.
Signal emerged from Whisper Systems, a startup co-founded by Moxie Marlinspike and Stuart Anderson that launched the RedPhone and TextSecure apps in May 2010. Twitter acquired Whisper Systems in November 2011, later releasing the code as open source. The nonprofit Signal Foundation was launched in February 2018 with $50 million from WhatsApp co-founder Brian Acton. As of January 2025, the platform had approximately 70 million monthly active users and had been downloaded more than 220 million times.
What this means for enterprise deployments
Signal is restricting screenshots on iOS 27 and Android to prevent OS-level AI agents from capturing private chats.