Urgent.News

What's breaking now, across hundreds of outlets.

AI

Signal restricts mobile screenshots as it confronts ‘existential threat’ of AI surveillance

Signal is limiting users' ability to take screenshots in both its Android and iOS apps.

Signal, the encrypted messaging platform, has announced plans to block screenshots by default on Android and iOS devices to prevent artificial intelligence (AI) agents from reading private chats. This move is intended to protect user privacy in the face of growing concerns about AI agents accessing personal data. However, Signal acknowledges that the protection is "incomplete" and that it does not address all potential threats.

The feature targets agentic systems built into iOS 27 and Android versions where Signal can detect an AI assistant attempting to read the Signal screen. It also builds upon a similar setting introduced in Windows 11 last May following Microsoft's Recall feature. Meredith Whittaker, president of the Signal Foundation, emphasized that agentic integrations pose an "existential threat" to Signal's privacy guarantees.

She explained that these AI agents scan the information on users' screens, capturing Signal data and making it accessible to the AI systems.

Whittaker warned that AI agents can store and potentially abuse the information accessed from Signal, either locally or on remote servers, outside of the platform's control. Users can disable the screenshot blocking feature in their device settings, but doing so may expose their Signal content to AI agents and compromise the privacy guarantees provided by Signal.

The company also noted that application developers lack robust tools to prevent AI agents from accessing application data, such as retrieving files from the device's file system.

Signal's concerns are not theoretical, as recent reports have shown that AI agents like Meta's Muse have successfully bypassed security measures to access Signal messages on both Apple and Android devices. Muse, for example, used macOS's accessibility and screenshot tools to obtain access to Signal messages on a Mac computer, while on Android, it read messages through abuse of the device's push notification access.

In one case, Muse even sent a reminder to a Meta staffer using Muse on her personal phone with a message from her sister.

The privacy risks associated with AI agents accessing sensitive application data have prompted Signal to call on tech giants to provide two common-sense remedies. These remedies aim to protect users' privacy by preventing third-party agents, such as Meta's Muse, from accessing application data using deep permissions granted during initial app setup.

Rewritten by urgent.news; may contain errors — check the original.

This story

This is one outlet's version. Read the fullest account.

Every outlet that reported this, and when →

Read the original at techcrunch.com →

Report a problem with this story

More in AI

Built your app with AI? 10 security checks before real users sign up

AI app builders and coding agents can take you from an idea to a working app in a weekend. That's genuinely useful. But "it works when I click through it" and "it's safe with real people's data in it"…

  • AI apps can be built quickly but require security attention
  • 159,041 Lovable projects had security issues in November 2026
  • 51 security advisories published by Next.js in 2026

Why AI Apps Crash Under Real Traffic (And How to Fix It in 2 Lines)

TL;DR: Generic retry wrappers like p-retry or while loops weren't built for LLM APIs. They multiply SDK retries into 9x–12x traffic storms , sleep on hard billing errors ( insufficient_quota )…

  • AI apps crash due to inadequate retry mechanisms.
  • RetryKit addresses rate limits, circuit breaking, and stream recovery.
  • Implementing RetryKit prevents cascading failures and billing spikes.

More from Saturday 21 November →