SIAgentNews.com
← All news

Signal blocks screenshots to prevent AI assistants from reading chats

The messaging app is disabling screenshot capabilities by default to stop AI agents from accessing private data via operating system accessibility features.

· First reported by TechTimes Vietnam

Blueprint drawing: picture for the story 'Signal blocks screenshots to prevent AI assistants from reading chats'
AI-made blueprint drawing for this story. No real people or products are shown.

The short version

  • Signal will block screenshots by default on Android and iOS to prevent AI assistants from reading on-screen data.
  • Users can manually re-enable screenshots in settings, but Signal warns this may expose private messages to third-party agents.
  • The move targets AI agents that use operating-system accessibility features to capture data from the screen.
  • Signal is calling on OS vendors like Apple and Google to provide better tools for limiting AI data access.

Why it matters

This move highlights a growing conflict between privacy-focused apps and the integration of AI into mobile operating systems. It signals a shift toward 'AI-aware' privacy standards where apps must actively block OS-level features to keep data secure.

The move against AI agents

Signal has announced a new security measure to prevent AI assistants from reading the content of private conversations. The messaging app will now block the ability to take screenshots on both Android and iOS devices by default. This change is designed to stop tools like Siri and Google’s Gemini from capturing on-screen text.

The company describes these AI agents as "hostile" because they can use built-in operating-system accessibility features to see what is happening on a user's screen. By blocking screenshots, Signal aims to ensure that the data remains private even when the app is open. The move is intended to keep the company's end-to-end encryption guarantees intact in an era of increasing AI integration.

Signal's president Meredith Whittaker stated that the move addresses "hostile" AI agents that can capture on-screen data. While the feature is disabled by default, the company is still allowing developers and users to restore the functionality through the app's settings. However, the company warns that choosing to turn this back on could leave private chats exposed to these automated systems.

How the technology works

To understand this change, it is important to look at how AI assistants interact with mobile hardware. Many of these assistants rely on accessibility features, which are tools designed to help users with disabilities navigate their phones. Because these features require high-level permissions to read the screen, AI agents can sometimes use them to "see" the content of apps like Signal.

Signal is specifically targeting the way these agents capture data through the operating system. When an app allows screenshots, it creates a pathway for an AI to potentially grab a visual record of a conversation. By blocking the screenshot function at the app level, Signal creates a barrier that prevents the OS from sharing that specific visual data with third-party agents or native assistants.

The company noted that this remedy is not a perfect solution for all types of data leaks. They acknowledged that sophisticated AI agents could still find ways to access information through the device's file system or deep accessibility settings. This means that while the screenshot block stops a common method of data capture, it does not completely seal the app from every possible technical exploit.

Previous actions and context

This is not the first time Signal has taken proactive steps to block OS-level features that threaten user privacy. The company previously implemented a similar block against Microsoft’s Recall feature. That specific tool was designed to capture and store what users saw on their screens, which posed a risk to secured chats.

The recurring theme for Signal is the tension between convenient AI features and the core principles of private communication. As AI becomes a native part of mobile operating systems, the company feels it must act to protect its users. They are currently highlighting a growing conflict where the very tools meant to help users might be used to undermine their privacy.

Signal's actions serve as a signal to the broader tech industry about the limitations of standard encryption. While encryption protects data while it is moving between devices, it does not necessarily protect data from the device's own operating system. The company is now forced to defend against the OS itself to maintain its privacy model.

Demands for OS vendors

Signal is not just changing its own code; it is also calling for broader changes from the companies that build mobile platforms. The firm is urging Apple, Google, and Microsoft to provide developers with clearer documentation and straightforward tools. They want a way to limit AI data access without making the user experience less functional.

The company specifically wants safeguards against third-party agents, such as Meta’s Muse, which could potentially access user data. They are pushing for a standard where developers can opt-out of AI data sharing more easily. This would allow apps to remain private without having to resort to aggressive blocks like the one currently being implemented.

If the major OS providers respond to these requests, it could lead to a new standard for AI-aware privacy. This would mean that apps could potentially have more granular controls over what parts of the screen or what types of data are shared with assistants. For now, however, Signal's manual block remains the primary defense for its users.

The future of app privacy

The situation described by TechTimes Vietnam suggests a broader industry challenge regarding application-level privacy. As AI assistants become more deeply integrated into the core of mobile phones, protecting user data becomes increasingly complex. This move by Signal underscores the fact that privacy is no longer just about the connection between two people, but about the relationship between the app and the device.

While Signal's screenshot block might only curb casual data leakage, it highlights the need for more sophisticated protections. It suggests that OS vendors will face increasing pressure to provide better APIs for developers. These APIs would allow apps to tell the operating system exactly what an AI assistant is allowed to see and what it is not.

The outcome of this conflict could shape the future of mobile platform policies. If Apple and Google provide better tools, other privacy-focused apps may follow Signal's lead and adopt these new standards. This could eventually lead to a mobile environment where AI can be useful without being a constant threat to the privacy of personal messages.

What to watch

  • How Apple and Google respond to Signal's request for better AI privacy tools.
  • Whether other privacy-focused apps implement similar screenshot blocks.

Try it

Check your Signal app settings to see if you can manually toggle the screenshot functionality on or off.

Sources and links

Read the original at TechTimes Vietnam

First reported by TechTimes Vietnam. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news