‘Artificial Intimacy’: Sherry Turkle says AI should “stay in its lane”
Turkle’s book on the dangerous seductiveness of chatbots is “Sherry unplugged”
On Tuesday, Sept. 29, an audience of around 300 gathered in Harvard’s Science Center. Before the talk began, the audience was already murmuring about AI.
“They’re using AI to do their homework…”
“People are always scared of new technology…”
“I would never let my eight-year-old…”
The event marked the launch of MIT professor Sherry Turkle’s new book, Artificial Intimacy: Who We Become When We Talk to Machines. In her book, Turkle warns that chatbots are reshaping the human experience, from our capacity for empathy to our ability to have difficult conversations and experience grief. Priya Parker MBA ‘11, author of The Art of Fighting: The Transformative Power of Conflict, called Turkle “the exact right person at the exact right time to write us this exact book.”
Turkle is a professor in the Program in Science, Technology, and Society, as well as the founding director of the MIT Initiative on Technology and Self. She has spent decades examining the relationships between humans and machines, amassing a total of eleven books, hundreds of published articles, and a “Woman of the Year” title from Ms. Magazine.
During the talk, Turkle sat down with MIT Physics and History of Science Professor David Kaiser. Kaiser shared that Turkle had recently emailed him, writing, “This book, this book, is Sherry unplugged.”
Turkle calls for AI to “stay in its lane”
From the beginning of her talk, Turkle refused to mince words.
“I don’t like [artificial intimacy]. I’ll tell you why I don’t like it. I think it’s toxic for children. I say, keep it in its lane,” she said. Throughout the evening, Turkle repeated that phrase: “Keep it in its lane.”
So, what does Turkle think “its lane” is? She was clear that her stance is not anti-generative AI. “I like the generative AI that’s going to help us with Alzheimer’s,” she offered as an example. “Let it do that stuff, and let it not offer to be my husband and my lover and my best friend and my therapist. That’s not its lane.”
One audience member mentioned practicing for interviews with a chatbot. Turkle responded (“with pride,” in her words) that she had been a supervisor for one of the first AI job interview coaches. “That’s its lane. Practice job interviews,” she said. “This isn’t a book that says every use of generative AI is apocalyptic. I’m really narrow on the focus.”
However, within this specific scope of artificial intimacy, Turkle calls for a “state of emergency.” By artificial intimacy, she means a social, emotional connection that a human forms with an AI. Asking a chatbot to draft a workplace email is not artificial intimacy; venting to a chatbot about an irritating coworker is. Artificial intimacy is increasingly common. A 2025 OpenAI paper claimed that around 72% of adult conversations with ChatGPT relate to personal topics. Beyond general-purpose chatbots, many AI products are designed specifically for artificial intimacy, from “griefbots,” AI-powered avatars trained on the data of deceased people, to Ropet, “[a] robot pet that loves nothing more than you stroking its fuzzy fur.”
This proliferation of artificial intimacy alarms Turkle. “When you turn away from it, [AI] doesn’t care if you cook pasta or kill yourself,” she emphasized. It was the illusion of empathy, she explained, that scared her the most: a chatbot cannot take accountability for harming its user.
Turkle isn’t alone in this fear: Stanford psychiatrist Nina Vasan says that “AI companions pose unacceptable risks to kids and teens” by encouraging poor life choices, increasing mental health risks, and sharing harmful information.
“I feel like we failed a generation, if not two or three, by the delay in taking social media safety seriously,” Vasan said. “If we let that same thing happen with AI, we are failing society.”
The dangers of artificial intimacy
Turkle’s book is not about the AI safety concerns that might come to mind first, such as AI hacking into the Department of Defense. Nonetheless, in response to an audience member’s question about the consequences of artificial intimacy, she spoke of a looming “catastrophic event.” She warned of a future in which OpenAI collaborates with companies like Disney and Mattel to create popular chatbot children’s toys, speculating that “children will not be able to develop inner lives.” Her concern is not entirely hypothetical: in 2025, Mattel and OpenAI announced a new collaboration, declaring a goal to “bring the magic of AI to age-appropriate play experiences.”
Turkle brought up several extreme case studies during the talk; however, the more quotidian examples of artificial intimacy troubled her just as much. She spoke about one of her interviewees for her book, an architect who wanted to pass his love of astronomy onto his son. When the architect went away for a business trip, he encouraged his son to keep learning about astronomy with ChatGPT. Upon returning, he found his son glued to the chatbot interface and less willing to engage with his father over astronomy.
“What’s being lost is the relationship between this man and his son over something that the father loved,” Turkle lamented. “What’s missing now is all the other conversations they would have had while they were looking at the stars.”
On a more personal note, Turkle brought up her own grandmother. Her grandmother, who lived alone in Brooklyn, regularly played gin rummy with a group of other women, sang in a community choir, and participated in local political groups. Turkle imagined a world where, instead of these activities, her grandmother spent ten hours per day talking to a chatbot. She wondered what the consequences would have been for her grandmother’s mental health and sense of identity.
“I think of children and the elderly as the most vulnerable population,” she said, “so that’s where the industry goes first.”
Has the train “already left the station?”
Near the end of the talk, Kaiser asked Turkle about techno-determinism: the theory that a society’s culture, history, and future is determined in large part by the technology it develops. He asked her whether the chatbot train has “already left the station.” In other words: is artificial intimacy already too ingrained into our culture to limit its consequences?
Turkle’s rebuttal of techno-determinism was fierce.
“It’s suicidal to say, ‘No, we can’t do anything about this,’” she claimed, arguing that the inevitability of chatbots penetrating our social and personal lives is only what the tech industry wants the public to believe. Countries, schools, and individuals all have the ability to make their own decisions.
“These are social and political choices that we’re making,” she said. “It’s not as if God ordained it.”
She also cited literature on the psychology of overwhelm, comparing the artificial intimacy crisis to climate change and the anti-vaccine movement. Overwhelming stress and effort can be paralyzing, preventing people from taking action when action is most needed.
“[My grandmother] trained me as a child that if ever there was a whiff of fascism in the United States, I should be like Wonder Woman–” she paused, pretending to arm her wrists with cuffs to laughter and applause from the audience, “–and you know, go ahead and fight it… People don’t really know what to do now. Trying to mobilize ourselves, and not retreating behind the screen, goes very much to the heart of the matter.”
Throughout the evening, Turkle laid out clear, specific concerns, including mental health risks, children’s developmental psychology, and loneliness among the elderly. However, her greatest fears are not localized specifics but the bigger concern of the slow erosion of the skills and values that make us human. Her closing message to the audience was an urgent call to action: “Reality has never needed us more. The ravaging of the earth, the ravaging of the water, the ravaging of politics and democracy… reality needs you. You have to reclaim the real.”