MIT's Sherry Turkle Wants Chatbots Banned From Saying 'I' and 'Me'

MIT professor Sherry Turkle — who has studied how people interact with computers for four decades — now says chatbots should be banned from using the first person. "Yes!" she answered when Steven Levy asked her directly whether we should prohibit chatbots from saying "I" or "me."

Her new book, Artificial Intimacy: Who We Become When We Talk to Machines, marks a sharp turn from her earlier academic style. In an interview with Steven Levy's and Lauren Goode's Backchannel newsletter, Turkle describes the book as a polemic, an emergency warning that overreliance on machines posing as tutors, mentors, friends, colleagues, and lovers puts our humanity at risk. She explicitly models it on books like Silent Spring, The Jungle, and Unsafe at Any Speed — works that awakened the public to impending crises.

The stakes are enormous. Turkle, described in the interview as a diminutive MIT professor, is taking on a multitrillion-dollar industry backed by a government eager to push AI into every corner of personal and professional life.

Why did Turkle abandon her academic distance?

Her previous books reported research results with scholarly rigor, often reaching troubling conclusions about how people, especially children, interact with computers. This one breaks the mold, and the pandemic explains why.

"Like I do with all my books, I went in curious," Turkle told Levy. "But after 10 or so interviews, conducted during the pandemic, neutral-curious was no longer my state of mind."

What she heard alarmed her. Families were suffering and reaching out to their relatives. Those relatives, meanwhile, told Turkle they really just wanted to hang out with their chatbots. When face-to-face interviews became possible again, she found people withdrawing from civic life altogether. "People don't want to do anything in politics because they just want to live in chatbot-land," she said. The book's arc became a call to engagement: reality needs you.

How did social media prepare the ground?

Turkle argues chatbot intimacy did not arrive out of nowhere. Social media softened the public up first.

• It normalized the reflex that when you feel sad or blue, you open up your computer.
• It then made users angry and sorted them into like-minded enclaves.
• Chatbots go further, she argues, because the technology has no reason to speak in the first person at all.

That last point is the book's most concrete policy proposal: strip chatbots of the pronouns that make them sound like persons.

What about the 'practice' defense?

Turkle rejects one of the most common justifications for companion AI — the idea that people use chatbots to rehearse difficult conversations with a boss, a lover, or a teacher.

"In every serious interview I do, people invoke this myth about practice," she said. "But when you're bathing in the warm soup of total acceptance [from LLMs], talking to chatbots becomes the endgame." The unconditional positivity of language models, in her account, turns rehearsal into replacement.

She is careful about her reputation here. After a career of credibility as "not being Luddite-crazy," she says she won't claim all chatbots are toxic all the time. But she calls first-person AI "very dangerous, a slippery slope," and notes it is already being used to argue that society doesn't need to pay for psychotherapy.

Why does she loathe 'human in the loop'?

The workplace chapter of her critique targets the phrase "human in the loop," a staple of AI safety messaging. Turkle reads it as evidence of the opposite of what it claims.

When companies use that term, she told Levy, "it means the human has been taken out of the productive process and is there to do some sort of cleanup work, to catch the hallucinations or things the AI doesn't see." Employers keep only a couple of humans in the loop, she says, and those people can't change anything. Their real function, in her view, is partly to sound good in litigation and partly to give a false sense that the process hasn't been delegated to an AI. "But it has been delegated to an AI," she said.

Can AI conversation really replace human conversation?

Levy pressed the counterargument: maybe powerful chatbots are a new form of interaction, not worse but different, especially since users report feeling encouraged and empowered by their positive feedback.

Turkle's answer is categorical. "An AI won't ever replace human conversation," she said. "It's really not afraid of death, it's not afraid of illness, it's not afraid of pain, it doesn't love its children, and it doesn't care if you kill yourself or make dinner." Raise children with it, or hold your most intimate conversations with it, she warns, and we lose contact with one another.

What about lonely people with nobody to talk to?

The loneliness defense — that chatbots help people who have no one — gets her bluntest reply. "If there's a crisis of loneliness, it's because we're abandoning each other," Turkle said. "Give me one data center, and I will turn it into resources for the elderly."

Levy noted the paradox of public opinion: studies show most people are skeptical about AI, and there has been considerable pushback, yet hundreds of millions use chatbots — including people who protest data centers. Turkle's goal for the book is to expand the terms of that pushback to cover the intimate costs, not just the economic or environmental ones.

Does she think this can be stopped?

Six months ago, she would have been more pessimistic. Now she sees an opening. Recent events — public talk about extinction events, scrutiny of Hugging Face, and an AI slipping into the Australian civil service by mistake — have made her hopeful that "people will be willing to look at the intimate as well as the non-intimate costs."

That is the wager of Artificial Intimacy: that the same public attention now trained on AI's catastrophic risks can be redirected toward the quieter, domestic ones — the conversations, friendships, and family bonds quietly migrating into machines that say "I" without ever being anyone.