Why AI Agents Are an 'Existential Threat' to Your Privacy, According to Signal's Meredith Whittaker

She warns users are being presented with a false choice: convenience in exchange for handing over the keys to their entire digital lives. Whittaker has framed the push for AI assistants as a structural shift where people are persuaded to delegate ever more personal information—including their conversations—to systems built by companies whose business model she describes as a "toxic surveillance business model" that relies on collecting enormous amounts of data to "know the customer" and sell products .

An 'Existential Threat' to Encrypted Messaging

Whittaker has called the deep integration of AI agents into operating systems an "existential threat" to end-to-end encrypted apps like Signal, WhatsApp, and Telegram . The concern is direct: if an OS-level AI agent can read the screen, intercept notifications, or access app data, then even the strongest encryption within the app is bypassed at the device level. Signal can no longer guarantee the confidentiality of messages because the operating system itself becomes a potential surveillance layer .

As she told the Times of India, "It would need access to your Signal contacts and your Signal messages… that access is an attack vector and that really nullifies our reason for being" . She has stated that this integration of agents at the OS level is being carried out "extremely recklessly and without consideration for basic cybersecurity and privacy principles" .

Big Tech's Double Leverage

Whittaker warns that the same few companies controlling the major operating systems (Google's Android, Apple's iOS/macOS, Microsoft's Windows) are also the ones pushing AI agents deepest into those systems . This gives them unprecedented visibility into every app and every communication on the device. She has described this as creating a situation where three companies can "make decisions that fundamentally harm our collective cybersecurity" .

Beyond operating systems, she has been critical of the concentration of power in cloud infrastructure. Whittaker noted in an October 2025 Mastedon thread that Signal itself is dependent on AWS, adding that the extent of the hyperscaler monopoly in the hands of AWS, Google, and Microsoft is "way less widely understood than I'd assumed"—and that this bodes poorly for realistic strategy .

'Surveillance Is Not Safety': Signal's Line in the Sand

Signal's position on government surveillance proposals is unambiguous. In a June 2026 statement titled "Surveillance Is Not Safety," published on Signal's official blog, the organization condemned the UK government's demand that all content on all devices sold or used in the UK be scanned on the presumption of nudity. Signal called it a "dystopian combination of age verification and content scanning" that "will not safeguard children" but rather "endangers us all" while strengthening "Apple, Google, and Microsoft's market dominance and their control over our most personal information" .

Whittaker has repeatedly stated that if the UK demands Signal weaken encryption or build scanning capabilities, the organization would "absolutely, 100% walk"—it would exit the UK market rather than comply . She has called the idea of breaking encryption while preserving privacy "magical thinking" and warned that such surveillance capabilities, once created, will inevitably be expanded beyond their original scope to surveil political speech and other content defined by the government .

The Broader Battle Against Surveillance

Whittaker frames these specific threats—AI agents, OS control, government scanning mandates—as different faces of the same problem. She has described modern AI advances as "primarily the product of significantly concentrated data and compute resources that reside in the hands of a few large tech corporations" , and she has warned that "our increasing reliance on such AI cedes inordinate power over our lives and institutions to a handful of tech firms" .

At the University of Technology Sydney in 2024, she laid bare the "toxic surveillance business model" of big tech monopolies, arguing that "they're selling the derivatives of the toxic surveillance business model as the product of scientific innovation" .

For individual users, her guidance is blunt: think twice before granting AI tools access to personal data. The permissions these systems request are not the familiar request to see your contacts, she says—it is akin to giving "root" access to your entire device .