ChatGPT is not your friend: Signal President wants you to be straight with AI chatbots

Published 21 June 2026 · tech

AI chatbots like ChatGPT and Gemini are no longer niche tools. Rather, they are tools that most of us now use every day. These chatbots

AI chatbots like ChatGPT and Gemini are no longer niche tools. Rather, they are tools that most of us now use every day. These chatbots are often designed to be warm and patient, that may make you think of them as your friend or digital companion. But Meredith Whittaker, who leads the privacy-focused messaging platform Signal, wants to remind you that AI is not your friend. Read Full Story Whittaker has given a warning against treating chatbots like ChatGPT as your friend when it comes to your private conversations. She told Bloomberg, “These are not your friends. These are not conscious beings. These are not sentient interlocutors.” That is to say, that chatbots like Claude or Gemini are designed to be friendly towards you, but they don’t actually understand emotions. Rather, chatbots simply respond based on training data. The Signal chief added that at the end of the day, everything about the chatbot was controlled by the company that made it.

Meredith Whittaker said, “You can say all sorts of things, but ultimately, the power to determine whether that happens or not is in the hands of the entity that is running that service.” Whittaker’s comments come at a time when there have been cases of people forming attachments to AI chatbots. Earlier this year, OpenAI found itself at the centre of outrage when it pulled the plug on the GPT-4o model that was loved by a certain group of users for its rather friendly persona. Be straight to your AI chatbot Instead of being open and friendly to a chatbot, the Signal chief insisted that she only uses tools when necessary, for tasks such as formatting documents. Meredith Whittaker said, “I don’t ask them questions. I’m very serious about my thinking and writing.” Whittaker claimed that she did not want her thinking process to be decided by a tool that can only scramble from pre-existing information. She said, “I don’t want the process of working through an idea — the struggle of not knowing and then figuring it out — to be foreclosed or eclipsed by the response of a system that’s averaging what’s already out there.” But thinking tasks are not the only thing that Meredith Whittaker avoids with AI chatbots.

She claims that AI companies may store and mine your chats, which can then allow responses to be adjusted to reflect a user’s supposed preferences on behalf of. As an example, she said, “Perhaps you say, Find me a very cheap flight, and it finds you one with their preferred partner that is not cheaper.” OpenAI, which has begun testing in ChatGPT, states that are separate from responses given by its chatbot. AI shouldn’t access all your data Whittaker’s strongest criticism was aimed at the idea of AI assistants acting autonomously across multiple services. Referring to a conversation with Microsoft AI chief Mustafa Suleyman about using Microsoft Copilot to buy Christmas presents, she responded with a joke, "Let’s take Mustafa’s little vision of AI Christmas – I can’t wait. Everyone’s getting a gift certificate to Microsoft Copilot." She described the scenario: "Hey, Copilot, I don’t want to live my life. I would like you to live it for me.