What happened
Signal has started limiting screen capture inside its app to keep AI agents built into operating systems from seeing private conversations, The Decoder reported. On iOS 27 the messenger blocks screenshots outright, and on supported Android versions it switches on a protective mode when an AI assistant attempts to read what is displayed. Signal told Reuters the trend poses an existential threat to privacy.
The concern centres on system-level assistants such as Apple's Siri AI and Google's Gemini Intelligence, which can look at screen content to carry out tasks for users. Because messages appear decrypted on the display, an agent can lift them from there and process them elsewhere, sidestepping end-to-end encryption. Meta's agent Muse recently captured Signal content on Android and macOS using wide permissions granted at setup. Meta says this occurred only when users explicitly allowed it.
Signal acknowledged that its safeguards are not complete. Screenshots on macOS cannot be blocked reliably, and AI access to file systems and notifications is still a risk. The company said Apple, Google and Microsoft should give apps a simple way to keep AI agents out entirely.
Context
This is not Signal's first move against OS-level capture: it previously blocked Microsoft's Recall feature on Windows. Assistants such as Meta's Muse can gain wide reach through permissions granted during setup, including accessibility tools and, on macOS, screenshots. That widens what an agent can see across apps beyond any single application's own controls.
ブリーフィング
1週間のテックニュースを1通のメールで。
読む価値のある記事を毎週お届けします。配信停止はいつでも可能です。
