Privacy and Technology Policies

Signal Automatically Starts Blocking Screenshots to Combat AI Agent "Espionage"

The encrypted messaging platform "Signal" has announced the default activation of a screenshot-blocking feature to prevent AI assistants like "Siri" and "Gemini" from reading user chats, describing the infiltration of AI into operating systems as an existential threat to privacy.

2026-11-20
4 min read
7 views
certi.news
Signal Automatically Starts Blocking Screenshots to Combat AI Agent "Espionage"

The encrypted messaging platform "Signal" has taken a decisive, proactive step against the infiltration of AI agents into mobile operating systems by announcing the default activation of a screenshot-blocking feature. This move comes as part of the platform's effort to protect chat privacy from built-in AI assistants that now possess the ability to "see" and analyze what is displayed on the screen.

Facing the "Existential Threat" to Privacy

Signal described the AI agents embedded in phones as posing an "existential threat" to the core of digital privacy. Under the new update, the platform will automatically block screenshots on the iOS 27 operating system, as well as on Android when AI assistants attempt to capture screen content for processing.

These measures directly target preventing tools like Apple's "Siri" and Google's "Gemini" from spying on chats. Although these assistants are designed to help users by reading screen content to provide more accurate answers, the side effect is that chats which are supposed to be end-to-end encrypted may end up stored on remote servers outside the control of the user or Signal itself.

New Protection Mechanisms on iOS and Android

The new privacy settings differ slightly between the two platforms, but the goal remains the same:

  • On iPhone: A new setting called "Block Siri AI" has been enabled by default.
  • On Android: Signal relies on the pre-existing "Screen security" option to disable the agents' ability to read.
  • Alerts: If an attempt to take a screenshot or an AI agent reading the content is detected, the user will see an alert giving them the option to disable the protection manually if they wish.

This move is not the first of its kind for Signal, as the platform previously blocked Microsoft's "Recall" feature last year from taking periodic snapshots of chats on Windows 11, reflecting the platform's ongoing strategy to build firewalls against operating system features that may violate privacy.

Technical Flaws and Existing Challenges

Signal acknowledged that this protection remains flawed due to restrictions imposed by operating system companies. If the other party in a chat disables the protection features on their device, Signal will not be able to prevent the other party's AI agent from reading the messages. Furthermore, these features do not prevent third-party agents, such as Meta's "Muse" agent, from reading messages via notifications or exploiting accessibility permissions that act as a "master key" to the system.

Signal strongly criticizes Apple and Google for failing to provide a simple programmatic way that allows developers to exempt their apps from AI access without making harsh compromises on the core user experience. On the "Mac" system, the platform still lacks any effective protection tools against these practices so far.

What Changes Practically for Users?

This development represents a shift in the concept of privacy; in the era of artificial intelligence, privacy is no longer just an individual setting, but rather a collective choice and a complex technical challenge. What Signal is doing today is an attempt to impose technical limits on operating systems that try to make everything available to artificial intelligence under the banner of "facilitating the user experience."

For the technical user, this means that relying on encryption alone is no longer sufficient if the operating system itself has the ability to photograph the screen and send data for cloud analysis. Signal emphasizes that it is holding discussions with Apple and Google to reach a system-wide unified standard that allows applications to formally and systematically block AI agents, and until then, privacy in the face of artificial intelligence remains an ongoing and incomplete battle.

News source
c
Author

certi.news

Explore this story

Related topics and entities

In the same category

You may also like

View all news