Signal introduces new restrictions to protect private messages
Signal messenger is introducing restrictions on taking screenshots in its mobile apps to protect users' private conversations from access by artificial intelligence. The new measures affect devices with iOS 27 and Android, but the protection mechanisms on these platforms will differ.
Signal Foundation President Meredith Whittaker announced this on November 20 in the company's official blog. Reuters also reports about the new restrictions.
The changes are due to the spread of AI assistants built directly into smartphone operating systems. Systems like Apple's Siri AI and Google's Gemini can analyze screen content to answer user questions and perform actions on their behalf.
Signal warns that such capabilities pose a privacy threat. Even if messages are protected by end-to-end encryption, an AI assistant could potentially read them after they are displayed on the device screen.
The information obtained may be stored in local databases or transmitted to remote servers used by AI systems. After that, Signal developers can no longer control further use of the conversation content.
“Currently, the integration of AI agents is changing this environment in ways that make it increasingly hostile to privacy. This is an existential threat to Signal and our ability to responsibly fulfill our mission,” — Whittaker stated.
What restrictions will appear on iPhone and Android
On devices with iOS 27, Signal blocks taking screenshots within the app. A new setting, Block Siri AI, is provided for this, which is enabled by default and prevents Siri's artificial intelligence from accessing the conversation content.
On Android, the mechanism works differently. The app automatically activates Screen Security when it detects an attempt by an AI assistant to access information on the screen.
Developers explain the differences due to Android's features. On some devices, it is impossible to reliably detect an attempt to take a screenshot if the corresponding function is blocked in advance. Therefore, Signal focused on detecting AI assistants' attempts to access screen content.
The new restrictions are enabled automatically, but users can disable the protection on their own in Signal settings if they need to take screenshots of conversations.
The company warns that disabling the function may give AI assistants access not only to the device owner's messages but also to their interlocutors' personal information.
On Apple computers, protection is limited for now
Signal acknowledges that the new measures do not ensure absolute privacy. Blocking screenshots does not exclude other ways for AI agents to obtain information, such as reading data already stored in the device's file system.
The situation with macOS remains especially difficult. According to developers, Apple does not yet provide sufficiently reliable tools for blocking screenshot creation by third-party AI systems.
Therefore, the company recommends that users of computers with macOS 27 manually add Signal to the list of excluded apps in Siri's system settings. This should limit the assistant's access to message content.
Signal already tried to solve a similar problem in 2025. At that time, the messenger blocked screenshot creation by default in its Windows 11 app to protect conversations from Microsoft's Recall feature.
Recall periodically saves screen images and allows artificial intelligence to analyze previously viewed information. Signal fears that such a mechanism could even save messages that are supposed to disappear automatically after being read.
The company called on Apple, Google, and Microsoft to provide app developers with tools to completely prohibit AI agents from accessing confidential information. This includes not only screen content but also notifications, files, and other app data.
Signal emphasizes that the problem affects not only messenger users. As AI assistant capabilities expand, similar risks may arise in banking apps, medical services, and other programs that process personal information.