Signal's Meredith Whittaker Warns: AI Chatbots Are Not Your Friends

Signal President Delivers a Stark Warning About AI Chatbots

In an era where millions of people turn to AI assistants for advice, companionship, and daily tasks, Signal President Meredith Whittaker is urging the public to pump the brakes. Her message is blunt and unambiguous: artificial intelligence chatbots are not conscious, they are not sentient, and they are absolutely not your friends.

Setting the Record Straight on AI Relationships

During a wide-ranging interview with Bloomberg covering privacy, policy, and Signal's role in the digital landscape, Whittaker was asked to weigh in on the privacy risks posed by popular AI chatbots such as ChatGPT and Claude. Her response left little room for interpretation.

"These are not your friends. These are not conscious beings. These are not sentient interlocutors," she stated plainly.

Whittaker admitted that she does occasionally use AI tools in her own workflow — primarily for basic formatting tasks — but draws a firm line when it comes to deeper engagement. She explained that she never uses these systems to work through ideas or answer meaningful questions, emphasizing, "I'm very serious about my thinking and writing, and I don't want the process of working through an idea to be foreclosed or eclipsed by the response of a system that's averaging what's already out there."

Her concern is not simply philosophical. It reflects a deeper understanding of how these systems function — as statistical engines drawing on existing data — rather than as thoughtful, independent thinkers.

The Hidden Dangers of AI-Powered Shopping Assistants

Whittaker also took direct aim at a recent prediction made by Microsoft AI CEO Mustafa Suleyman, who suggested that users could soon rely on Microsoft Copilot to handle their holiday shopping entirely. In Suleyman's vision, Copilot would monitor family group chats, identify gift preferences, and make purchases autonomously on the user's behalf.

For Whittaker, this scenario represents something far more troubling than a convenient holiday shortcut.

"What you've just described is a system with very pervasive access across multiple applications and services," she warned. To function as Suleyman described, such a system would require access to a user's credit card information, web browser activity, private messaging platforms like Signal, the ability to send messages impersonating the user, their home address, and their personal calendar.

A Backdoor by Another Name

From a security standpoint, Whittaker argued that granting an AI assistant this level of access would effectively create a backdoor into platforms like Signal — a direct contradiction of everything encrypted messaging is designed to prevent.

"In the context of Signal, it would constitute a kind of a backdoor," she said, underlining the serious security implications that tend to get buried beneath the convenience-focused marketing of AI products.

Why This Warning Matters

Whittaker's comments arrive at a critical moment, as AI companies race to deepen their integration into everyday life. While the technology undeniably offers real utility, her perspective serves as an important counterweight to the hype.

Users who treat AI chatbots as trusted confidants — sharing personal struggles, financial details, or sensitive communications — may be exposing themselves to significant privacy risks without fully understanding the consequences. Every conversation fed into these systems contributes to data collection at a scale most people rarely consider.

For privacy advocates and everyday users alike, Whittaker's message is a timely reminder: convenience should never come at the cost of security, and knowing what you are actually talking to matters more than ever.