SIAgentNews.com
← All news

Signal restricts screenshots to block AI agents from reading chats

The encrypted messaging app is introducing new blocks on screen captures to prevent device-based AI from accessing private conversations.

· First reported by livemint.com

The short version

  • Signal will block screenshot access on phones and turn these restrictions on by default.
  • The move aims to stop AI agents from pulling data out of Signal's secure, end-to-end encrypted environment.
  • Signal called the rise of agentic AI an 'existential threat' to its privacy guarantees.
  • The company is urging Apple, Google, and Microsoft to provide developers with better ways to opt out of AI data processing.

Why it matters

This highlights a growing conflict between privacy-first apps and the push for 'agentic' AI that can perform tasks by reading your screen. It signals a shift toward a more restrictive mobile environment where apps must actively fight to keep data private from the operating system.

The threat of agentic AI

Signal is taking a stand against the growing trend of 'agentic' AI. These are AI systems designed to act on behalf of users by finding information across different apps and performing tasks like booking appointments or answering emails without direct instructions.

Because these agents need to see what is on a user's screen to work, they can potentially bypass the security measures of individual apps. Signal warns that this capability allows AI to pull private messages out of its secure environment and onto other servers.

The company describes this shift as an existential threat to its mission. It argues that the push to embed AI into consumer devices is eroding the ability of privacy-focused apps to guarantee user confidentiality.

New screenshot protections

To counter these risks, Signal is introducing new settings to block screenshot access on mobile phones. These restrictions will be turned on by default to prevent AI assistants from scanning private chats.

Users will see pop-up windows that offer options to disable these blocks if they detect a screenshot attempt. This system is designed to trigger when the app notices an AI assistant trying to access the screen or a user manually trying to capture the content.

Signal previously implemented similar blocks for Microsoft's Recall AI on Windows systems last year. However, the company noted it is currently unable to offer the same reliable protections on Apple's macOS platform.

Bypassing app security

The need for these protections follows reports of AI agents successfully bypassing official tools to access private data. For example, Meta's Muse agent was reportedly able to access Signal messages on both Android and Apple devices by using broad operating system permissions.

On macOS, the agent reportedly used accessibility and screenshot tools intended for people with disabilities to navigate to Signal. It then took screenshots of chats, extracted the text from those images, and reproduced the messages inside its own app.

While Apple recently updated its software to close that specific pathway, it introduced new permissions for agents with similar capabilities. Signal argues that developers need straightforward ways to opt out of screen reading, notifications, and file system access entirely.

A call for tech giants

Signal is calling on the major operating system providers—Apple, Google, and Microsoft—to provide more concrete tools for developers. They want specific ways to block AI agents from accessing sensitive data rather than relying on individual app workarounds.

The company also wants these tech giants to be more transparent about how their AI agents function. Providing more information would allow developers to assess exactly what data the bots are accessing so they can adjust their systems accordingly.

Signal emphasizes that because operating systems maintain the environment where apps exist, they hold the power to shape the privacy landscape. The company wants a system where apps can choose to remain private from the AI tools built into the device hardware and software.

The privacy trade-off

Signal's core business model relies on limiting the data it stores on its own servers. By using end-to-end encryption, the service ensures that it cannot access or read the conversations conducted by its users.

The company argues that AI systems scanning screens undermine these protections by moving data into environments where Signal has no control. This creates a situation where the security of a message depends on the permissions granted to the operating system rather than just the app itself.

As the tech industry races to roll out more capable AI agents, Signal's move marks a significant escalation in the fight for data sovereignty. It highlights the growing tension between the convenience of autonomous AI and the fundamental right to private communication.

try_it/watch_list_placeholder_not_used_as_per_instructions_but_included_for_structure_compliance_if_required_by_schema_logic_none_provided_in_source_for_specific_actions_or_watch_items_except_general_context._Note_The_prompt_asks_for_these_fields_but_the_source_does_not_provide_specific_actions_or_watch_items_beyond_the_news_itself._I_will_provide_the_JSON_as_requested_with_empty_or_relevant_content_if_possible.

try_it_field_empty_as_source_does_not_provide_specific_action_watch_field_contains_questions_about_future_os_updates

What to watch

  • How will Apple and Google respond to Signal's request for better opt-out tools?
  • Will other encrypted apps adopt similar screenshot blocking by default?

Sources and links

First reported by livemint.com. 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