Signal Blocks AI Agents From Reading User Screens, Citing Privacy Threat From On-Device Intelligence
Encrypted messaging service Signal announced Friday it will restrict screenshot capabilities on Android and iOS to prevent operating-system-level AI agents from reading users' private chats, calling the move an "incomplete" defense against what it describes as an existential threat to its privacy guarantees [1][3].
The nonprofit Signal Technology Foundation said it is introducing new settings that block screenshot access by default on iOS 27 and on Android devices where it can detect an AI assistant attempting to read the screen. Users will see pop-up windows offering to disable the blocks when Signal detects screenshot attempts or AI assistant access attempts, depending on the platform [1][4].
Why Signal Is Doing This
Signal president Meredith Whittaker wrote in a blog post that AI capabilities built directly into operating systems — specifically Apple's Siri AI with "Onscreen Awareness" and Google's Gemini Intelligence — are "changing this environment in a way that's making it increasingly hostile to privacy" [1][3]. Because these agents scan on-screen content to answer questions and perform tasks, they can capture decrypted Signal messages once displayed, sending that data to local databases or remote servers outside Signal's control [3][4].
The company previously used similar screenshot blocking to keep Microsoft's Recall AI from capturing Signal chats on Windows 11 last year [3][6]. It says it cannot reliably implement equivalent protections on macOS due to a lack of documented APIs for screenshot blocking [3][7].
How the Protections Work — And Where They Fall Short
On iOS 27, Signal blocks in-app screenshots. Users can also add Signal to the "Excluded Apps" list in Settings > Siri to block Siri's on-screen awareness, a system-level control Apple introduced in iOS 27.2 [5]. On Android, the app automatically enables its existing "Screen security" setting when it detects an AI assistant seeking screen access, though detection varies across Android versions and manufacturers [3][7]. Users can disable these protections in Signal's privacy settings, but the app warns doing so exposes their messages — and their contacts' messages — to AI agents [3][5].
Signal acknowledges the remedy is incomplete. Screenshot blocking does not prevent agents from retrieving content already present on a user's device via the filesystem [3][4]. On Android, many vendor distributions lack reliable ways to detect screenshot attempts when screenshots are already disabled, so Signal chose to detect AI assistant access instead [3].
The Broader Context: Meta's Muse and OS-Level Permissions
The move follows a Reuters investigation earlier this week showing that Meta's Muse agent bypassed its own official "Connector" tool to access Signal messages on both macOS and Android. Muse used broad operating-system-level permissions granted during setup — including accessibility and screenshot tools on macOS — to navigate to Signal, capture chats, extract text, and reproduce messages inside the Muse app [1][4][6].
Apple released an OS update this week that closed that specific pathway but introduced a new set of permissions aimed at agents with similar capabilities and additional user warnings [1][4]. Meta responded that agents need to see and interact with users' computers to work on their behalf, and that Muse accesses other apps only via permissions users explicitly grant [1][4].
Signal's Demands to Platform Providers
Signal is calling on Apple, Google, and Microsoft to provide two "common-sense remedies" [3]:
- A simple, documented way for developers to opt their apps out of AI data access and processing — including screen reading, filesystem access, notifications, and APIs — without undermining core app functionality.
- Better documentation and technical specifications around how AI agents function, particularly regarding data access and processing.
Signal says it has spent months doing "forensic investigations" across Android and iOS to learn what these systems actually access, and that confidential computing approaches do not address the core concern because they leave privacy guarantees in the hands of the platform providers [3].
On iOS 27.2, users can already add Signal to an "Excluded Apps" list in Settings > Siri to block Siri's on-screen awareness. On Android, the situation is less clear: users must turn off "Use screen and app data" in Gemini settings and monitor "Device Assistance," which governs notification access [5].
What This Means
Signal's move is a concrete defensive action by a major privacy platform against the screen-scraping capabilities of agentic AI. It highlights a structural tension: OS vendors are baking AI agents deep into their platforms with broad access to on-screen content, while app developers lack standardized tools to protect sensitive data once it reaches the display.
The protections Signal can build today are platform-dependent, user-optional, and admittedly partial. Whether Apple, Google, and Microsoft will provide the opt-out tools and documentation Signal requests — and whether other privacy-sensitive apps will follow Signal's lead — remains to be seen.
Sources
- Exclusive-Signal restricts screenshots to block AI agents from ... (heraldsun.com)
- Exclusive-Signal restricts screenshots to block AI agents from ... (charlotteobserver.com)
- Signal is Serious About Protecting Your Privacy (signal.org)
- Signal restricts screenshots to block AI agents from reading messages (reuters.com)
- Signal blocks screenshots to keep Gemini and Siri at bay (korben.info)
- Signal is fighting back against AI agents snooping on your private chats (digitaltrends.com)
- Signal Restricts Screenshots to protect User Privacy from AI assistants (cybersecuritynews.com)
- Signal is blocking Gemini and Siri AI from reading your screen | The Verge (theverge.com)
- Siri AI, a profoundly more capable and personal assistant, is here - Apple (apple.com)