Signal, the encrypted messaging application, has begun blocking Google Gemini, Apple Siri, and other AI agents from accessing user messages unless explicit permission is granted, according to a blog post published this week by Signal CEO Meredith Whittaker. The company took the step because agentic systems built into iOS and Android can capture private communications and potentially send them to external servers, undermining the app's core privacy protections. Signal criticized Apple, Google, and Microsoft for failing to give developers adequate tools to prevent AI agents from accessing application data.
Updates to Signal's iOS and Android apps now include controls that by default prevent Siri on iOS and Gemini and other agents on Android from capturing screenshots, which AI agents process using optical character recognition to extract text from the screen. Users who want to take screenshots of their conversations can turn off this protection through Signal's settings menu, though the company warns that doing so risks exposing Signal content to AI agents. The blocking measures aren't available on macOS yet because Apple and other platform owners don't provide ways for developers to restrict AI data access without compromising core user experience, according to Whittaker. Apple is introducing more granular controls for two APIs—Device Control and Data Access, and Full Disk Access—in macOS 27.3, currently in public beta testing, though these changes appear focused on limiting third-party access rather than restricting Apple's own Siri service.
Whittaker explained that Google and Apple are embedding agentic capabilities like Gemini Intelligence with Google Gemini and Onscreen Awareness with Siri AI deeply into their operating systems, and these OS AI agents "scan" information on the user's screen to gain access to personal data. She acknowledged that screenshot blocking is only a partial solution since AI agents may have other methods to access application data through the device file system. "It is a serious problem that Signal—and application developers beyond—lack robust tools to prevent AI agents from undermining application-level privacy," Whittaker stated.
The tension between AI agents and privacy reflects a broader backlash across digital commerce and finance. Amazon has blocked shopping agents like Meta Muse from its e-commerce site, while Walmart has invited third-party agents to interact through its AI Gateway. United Airlines has banned transactions from unauthorized AI bots, and similar barriers have been erected by Target, Home Depot, Etsy, eBay, JPMorgan Chase, Charles Schwab, Citigroup, and Wells Fargo. Anthropic has blocked subscription customers from using third-party harnesses. A growing industry in bot detection and mitigation, led by companies like Cloudflare, Akamai, and DataDome, is emerging to address these concerns. As AI agents become increasingly capable and autonomous, the risks associated with this level of access will grow substantially, according to Apple's statement when introducing the Full Disk Access change.
After a year or two of agent vendors arguing for access to all digital data for personalization and context, the report notes the backlash has arrived. AI agents are supposed to serve the user, but when they send data to the cloud and come between customers and merchants, they wear out their welcome. The question people are now asking is whether their agents really serve them or someone else. Platform owners will face mounting pressure to choose between promoting their own agentic products and providing developers with meaningful privacy safeguards—a choice that may ultimately determine whether users trust the next generation of intelligent assistants or reject them entirely.

