Opinion Anthropic Mathematician Claims Claude Helped Prove the Riemann Hypothesis — in a 17-Million-Line Lean Formalization Nobody Has Verified Security Signal Blocks Siri and Gemini From Reading Your Chats, Calling OS-Level AI Agents an Existential Threat Apple Intelligence A Removed Apple Job Listing Reveals a Mac AI Agent That Browses, Codes and Controls Apps — Targeted for Late 2027 Chatbots Mira Murati's Thinking Machines Releases Inkling 2 Mini, a 26.4B Looped MoE That Runs on a 32GB Mac Business IEA Says Data Center Power Demand Could Double to 1,000 TWh by 2030, With AI Supplying Three Quarters of the Growth Chatbots ChatGPT Now Transcribes as You Talk: OpenAI Removes the 10-Minute Cap and Streams Words Live on Every Plan Business Ex-Alibaba Executive's Everful Raises $48.5 Million to Put AI Agents in Charge of Sourcing From 30,000 Factories AI Agents Google Built a Mirror Dimension of the Internet to Train Its Chrome AI Agents Without Breaking the Real Web Opinion Anthropic Mathematician Claims Claude Helped Prove the Riemann Hypothesis — in a 17-Million-Line Lean Formalization Nobody Has Verified Security Signal Blocks Siri and Gemini From Reading Your Chats, Calling OS-Level AI Agents an Existential Threat Apple Intelligence A Removed Apple Job Listing Reveals a Mac AI Agent That Browses, Codes and Controls Apps — Targeted for Late 2027 Chatbots Mira Murati's Thinking Machines Releases Inkling 2 Mini, a 26.4B Looped MoE That Runs on a 32GB Mac Business IEA Says Data Center Power Demand Could Double to 1,000 TWh by 2030, With AI Supplying Three Quarters of the Growth Chatbots ChatGPT Now Transcribes as You Talk: OpenAI Removes the 10-Minute Cap and Streams Words Live on Every Plan Business Ex-Alibaba Executive's Everful Raises $48.5 Million to Put AI Agents in Charge of Sourcing From 30,000 Factories AI Agents Google Built a Mirror Dimension of the Internet to Train Its Chrome AI Agents Without Breaking the Real Web

Signal Blocks Siri and Gemini From Reading Your Chats, Calling OS-Level AI Agents an Existential Threat

Signal Blocks Siri and Gemini From Reading Your Chats, Calling OS-Level AI Agents an Existential Threat

Signal announced new protections on November 20 that block screenshots by default on iOS 27 and auto-enable Screen security on Android when an AI assistant tries to read the screen, aiming to stop Siri AI and Gemini Intelligence from capturing private messages. President Meredith Whittaker called OS-level agents an existential threat and asked Apple, Google and Microsoft for a documented opt-out mechanism.

Signal, the encrypted messaging app, declared war on the AI assistants now built into phone operating systems. In a blog post published November 20, Signal president Meredith Whittaker announced new defaults that block Siri AI on iPhones and Google's Gemini Intelligence on Android from capturing Signal conversations through screenshots — and described the situation as an existential threat to the app's core promise.

The mechanism is the weak point, not the encryption. Features like Siri AI's Onscreen Awareness and Gemini Intelligence work by scanning what is on the screen, so an agent answering "what did my lawyer say" captures the chat in the process. Once captured, Whittaker wrote, the data is made available to whatever local database or remote server the OS-level agent is using at that moment — outside Signal's control, and exposed for every participant in the conversation, not just the person holding the phone.

The changes differ by platform. On iOS 27, a new Block Siri AI setting is on by default: it prevents Siri from scanning Signal content and, as a side effect, blocks screenshots inside the app and the App Switcher. On Android, where Signal cannot reliably detect screenshot attempts across manufacturers, the app watches for an AI assistant trying to read the screen and automatically enables its existing Screen security setting when it sees one — with the caveat that this only works on devices where detection is possible. On macOS, Signal says it currently has no reliable way to block capture at all, and can only advise users to add Signal to the Siri Excluded Apps list.

Users can switch the protections off in settings, but Signal warns the consequences extend beyond their own chats: doing so exposes their contacts' messages too, and breaks the app's privacy guarantees for the whole conversation. The move repeats a play Signal ran in 2025, when it blocked screenshots on Windows 11 by default to keep Microsoft's Recall feature from archiving chats.

Whittaker is upfront that the fix is incomplete. Blocking screenshots does nothing about agents reading content from the device filesystem, and third-party agents such as Meta's Muse can use accessibility permissions — which she likens to a skeleton key — to see what is on screen. Signal says it has spent considerable developer time on forensic investigations across Android and iOS trying to determine what actually happens when OS AI systems access application data, and that Apple, Google and Microsoft provide no way for developers to restrict AI data access without compromising core functionality.

The ask, accordingly, is structural: Signal wants the operating-system makers to ship an easy, documented way for apps to opt out of AI data access across screen reading, filesystem access, notifications and APIs, plus clearer documentation of how their agents handle app data. It says it is talking with Apple and Google and is cautiously optimistic. The broader standoff is worth watching: OS owners are simultaneously building their own agents and policing third-party ones, and app developers are discovering that privacy guarantees made at the application layer can be silently overridden one layer below.

Comments (0)

Log in to join the discussion

Log In

No comments yet