
Signal President Meredith Whittaker Warns AI Chatbots Are Not Your Friends
Key Takeaways
- Whittaker says AI chatbots are not sentient, not friends, and should not be trusted confidants.
- Her warning emerged during a Bloomberg interview on policy, privacy, and secure messaging.
- She uses AI tools sparingly for formatting and avoids asking them questions.
Whittaker’s AI warning
Signal President Meredith Whittaker told Bloomberg, as carried by TechCrunch, that “These are not your friends. These are not conscious beings. These are not sentient interlocutors.”
“Signal President Meredith Whittaker has issued a stark reminder to users of artificial intelligence chatbots: these systems are not sentient beings and should not be treated as confidants”
In the same Bloomberg interview discussed by TechCrunch, Whittaker said she uses AI tools “to format a document here and there,” but added, “I don’t ask them questions.”

Whittaker framed the concern as a threat to how people think and write, saying she does not 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.”
TechCrunch also reported Whittaker’s view that AI systems with broad access would create a “backdoor” into private communications in the context of Signal.
Copilot and the backdoor
Whittaker’s warning centered on a Microsoft Copilot scenario described by TechCrunch, where Copilot would eavesdrop on a family group chat to determine who wants what for Christmas shopping.
TechCrunch reported her argument that this would require giving Copilot “access to my credit card, my browser, my Signal, the ability to message my siblings on my behalf, my home address [and] my calendar.”

In the same framing, she said “What you’ve just described is a system with very pervasive access across multiple applications and services,” and that, specifically for Signal, it would be “a kind of a backdoor.”
Zamin.uz and Bitcoin World also relayed the same core claim from Whittaker that such systems are not conscious beings, while emphasizing the privacy risk of entrusting AI with personal data.
Privacy default vs. AI
While Whittaker pressed for boundaries around AI assistants, dev.ua described Signal creator Moxie Marlinspike launching the AI assistant Confer and positioning it as an alternative to traditional chatbots where privacy is often manually enabled.
“According to TechCrunch, Signal President Meredith Whittaker said, "These are not your friends”
dev.ua said the developers intend the most private mode to be basic and “not allow even server administrators to read conversations,” and it described Confer encrypting user interactions before they reach the server with access keys stored only on the device.
The same dev.ua account said Confer operates in a trusted execution environment where data remains inside a secure “container,” and it claimed it doesn’t store chats for further verification, model training, or sharing with third parties.
In contrast to Whittaker’s warnings about cross-application access, dev.ua said Confer uses remote attestation so a user can verify that the declared code is running and that each release is digitally signed with a published description of the entire software stack.
More on Technology and Science

Mukesh Ambani Announces Reliance $110 Billion AI Investment in New Delhi and Jamnagar
15 sources compared
US Commerce Secretary Howard Lutnick Pressures ASML Over Possible EUV Machine in China
16 sources compared

Saudi Doctors Successfully Separate Filipino Conjoined Twins Olivia and Gianna in Riyadh
11 sources compared
FDA Advisers Recommend Approving Moderna’s mFlusiva mRNA Flu Vaccine for Older Americans
11 sources compared