Signal is now blocking screenshots by default on iOS 27 and restricting on-device AI assistants from reading the app on Android, a direct countermeasure against the AI agents that Apple and Google have woven into their mobile operating systems. Signal president Meredith Whittaker announced the change in a blog post published on November 20 titled "Signal is Serious About Protecting Your Privacy," and did not mince words about the stakes: operating-system-level AI, she wrote, is "an existential threat to Signal, and our ability to responsibly uphold our mission."
The move, first reported as an exclusive by Reuters, is the sharpest escalation yet in the quiet conflict between privacy-first app makers and the platforms they depend on — and a pivotal moment for anyone following the latest AI developments in consumer technology. Signal's encrypted messaging app has long promised that only you and the person you're talking to can read your messages. Its argument is that agents built into iOS, Android and Windows now break that promise from the outside, by reading the screen itself.
What Signal Is Actually Blocking
The protections differ by platform because the underlying tools do. On iOS 27, Signal now prevents screenshots inside the app by default. On Android, the app automatically enables its "Screen security" setting when it detects an AI assistant attempting to access the screen — a deliberate design choice, Whittaker explained, because many Android vendors outside Google's Pixel line offer no reliable way to detect a screenshot attempt once screenshots are disabled. Rather than leave users with screenshots that mysteriously stop working, Signal targets the agent activity itself.
Users who want the ability to take screenshots can switch the protection off in Signal's settings. The company is explicit about the trade-off: doing so could expose both your messages and your contacts' messages to whatever AI agents are watching the screen, breaking the privacy guarantee Signal extends to every participant in a conversation.
The Recall Precedent
This is not the first time Signal has drawn a line against operating-system AI. Last year, the app used the same screenshot-blocking mechanism to prevent Microsoft's Recall AI from capturing Signal messages on Windows 11 by default. According to Signal's own technical footnotes, Recall captures the screen every few seconds, extracts text from those screenshots using optical character recognition, and populates a searchable on-device database — a process that could expose and retain Signal data indefinitely, including disappearing messages meant to vanish.
Why OS-Level AI Alarms Signal
In the blog post, Whittaker frames the issue structurally rather than as a dispute with any single company. Google, Apple and Microsoft produce the three dominant consumer operating systems — "the oceans in which Signal, and other applications, must swim," as she puts it, systems that "dictate the conditions of life for the application ecosystem."
The problem is how agents operate. To answer questions and complete tasks, OS-level AI scans what is on the screen. Anything visible in Signal at that moment — message content, contact identities — is captured and made available to the agent, potentially landing in a local database or a remote server outside Signal's control. The company's end-to-end encryption guarantees the transport of a message, not its appearance on a screen that the operating system is photographing. Whittaker writes that these integrations are "making it increasingly hostile to privacy" for the applications swimming in that environment.
The Gaps Signal Can't Close Yet
Signal is candid that the remedy is incomplete. Coverage is uneven across Android's many distributions. Screenshot blocking, moreover, does not address what Whittaker calls "more sneaky ways AI agents could potentially access application data," such as retrieving content already stored on the device's filesystem.
The most glaring hole is macOS. Signal says it cannot reliably block screenshots there at all, citing a lack of documented APIs — Apple has deprecated older approaches in favor of screen-protection solutions that only work for streaming media and DRM-protected content, which Whittaker describes as "a minefield of bewildering decisions that place user privacy at risk." For macOS 27 users, the company's only interim advice is to manually add Signal to the Excluded Apps section under System Settings > Siri, which stops proactive access by Siri AI but not screenshots.
The Documentation Fight
Underlying the technical measures is a grievance about opacity. Signal says it has spent months of developer time conducting forensic investigations across Android and iOS to understand what actually happens when AI systems access application data — because Apple, Google and Microsoft provide little documentation about how their agents work.
The company is calling on the three OS vendors to supply two things: accurate technical specifications describing how agents access and process data, and simple tools that let developers restrict that access. It also flags third-party agents such as Meta's Muse, which it says leverage accessibility permissions granted during setup — permissions "meant to empower actual people" — as a "skeleton key" for AI data access. For now, Signal says it is "cautiously optimistic" that Google and Apple are engaging with the company on how these systems should interface with privacy-sensitive apps.
The Canary in the Coal Mine
The announcement drew immediate coverage from Reuters, CNET, Engadget, The Decoder and Fast Company, and lands amid broader scrutiny of agentic AI's data habits. Reporters previously documented Meta's Muse agent accessing Signal messages during internal employee testing and mixing up conversations — an episode that foreshadowed exactly the risk Signal is now engineering against.
Whittaker's closing argument is that this fight extends well beyond one app. "Signal is the canary in the coal mine," she wrote, "but this affects the entire ecosystem and should galvanize anyone who cares about these fundamental rights." Whether Apple, Google and Microsoft respond with real tools — or with documentation that merely explains what their agents already do — will shape what privacy means on the devices most people carry.
Stay Ahead of AI
Get the latest AI news, analysis, and breakthroughs — all in one place. Follow our artificial intelligence updates coverage. Read more AI news →.
