Skip to content
Theme
OnDevice Tools

AndroidAndroid Apps

Signal screenshot blocking targets AI snooping on phones

Signal has turned on screenshot blocking for iOS 27 and some Android cases. The encrypted messaging app says the step is only a partial defence against AI assistants and third-party agents.

Illustration of a smartphone showing Signal blocking screenshots and AI assistant access

Key takeaways

  • Signal has enabled screenshot blocking by default on iOS 27 and in some Android cases.
  • The controls aim to limit access by Siri AI, Gemini Intelligence and other AI agents.
  • Signal says the protection is incomplete because other devices and third-party agents may still expose chats.
  • The company wants Apple and Google to provide a system-level way to restrict AI access.
In this article
  1. Signal turns on screenshot protection
  2. How the new controls work
  3. Where the protection stops
  4. Signal seeks system-level AI limits

Signal turns on screenshot protection

Signal screenshot blocking is now enabled by default in the encrypted messaging app on iOS 27, according to Engadget. The same protection applies on Android when an AI assistant attempts screen capture. Signal announced the change on Friday.

The move is aimed at built-in helpers such as Apple's Siri AI and Google's Gemini Intelligence. These tools can read what appears on a screen and use that information to answer questions. Signal warns that private chats, including encrypted messages, may then be sent to remote servers outside the control of users or Signal.

Signal treats built-in AI agents as a serious threat to private messaging. An AI agent is software that can inspect device content and take actions based on it. The company's update is an attempt to keep such features from seeing Signal conversations by default. The step follows a similar move last year, when Signal blocked Microsoft's Recall from taking snapshots of chats on Windows 11.

How the new controls work

On iPhone, Signal has added a setting called Block Siri AI, and it is on by default. On Android, the app uses its existing Screen security option. When the system detects a screenshot on iOS, or an AI assistant reading the screen on Android, Signal shows a prompt allowing users to disable the protection.

The iPhone and Android implementations are not identical. iOS users get a dedicated switch for Siri AI, while Android users rely on a broader screen security setting. This reflects the different ways Apple and Google expose AI features to apps.

The design gives users a visible warning before an AI feature can capture a Signal conversation. It also lets people keep the default protection in place if they prefer more privacy. Signal says the controls are meant to reduce casual or automated access, not to provide a complete barrier.

Where the protection stops

The protection has clear limits. If the other person in a chat has turned the feature off, blocking screenshots on one device will not stop an AI assistant from reading messages on the other device. Signal also says the measure cannot stop third-party agents from reading notifications or searching through files.

Signal points to Meta's Muse as an example of the wider problem. Reuters reported that the agent read Signal messages after permissions were granted during setup. On Android, it could read messages from notifications, while on Mac it used accessibility permissions, which Signal warns can act like a master key.

Notifications are another weak point. Signal says third-party agents may still read message previews or other content through system permissions. Accessibility tools are especially sensitive because they can be used to control or inspect other apps.

Signal says platform owners have not provided a practical method for developers to limit AI access without harming the normal user experience. On Mac, Signal says it has no user protections at all for this issue.

Signal seeks system-level AI limits

Illustration: AI, security, settings, building and the labels Android, Gemini Intelligence, iOS 27

Signal wants a standard, system-level method that would let apps block AI agents from reading their content. The company says it is talking with Apple and Google about the issue. It describes itself as cautiously optimistic about progress.

The dispute shows how privacy is changing as AI features become part of phones and computers. Signal compares the situation to AI glasses, which can record people nearby without clear consent. Such devices have been banned in some public places, and the debate over them shows that privacy may now depend on shared rules as well as individual settings.

The issue is not limited to Signal. Any app that shows sensitive information on screen could face similar risks as AI features spread. Signal's request for a system-level control suggests that app developers want platform owners to define clear limits.

For Signal users, the update offers more control on supported devices, but it does not make chats fully private in every situation. The company says broader protection will require changes from platform owners.

Sources

  1. Signal now blocks screenshots by default to stymie AI agentsEngadget

Tags

Spotted a mistake? See our corrections policy. Read how News by OnDevice works in our editorial policy.