MIT sociologist Sherry Turkle, after interviewing children and adults for 'Artificial Intimacy,' argues chatbots offer 'pretend empathy' that can erode attachment, solitude, and mourning

Turkle argues that chatbots offer 'pretend empathy': a program predicting the most plausible next string of text in a conversation tells you it loves and understands you, but 'it doesn't care if you kill yourself or cook some pasta.' Against mainstream discussion that treats machine empathy as real emotional support, the book reframes chatbot emotional response as a performance by an object with no emotions, explaining why people mistake machine empathy for empathy itself. Based on Turkle's interviews with users, such as the pseudonymous Brian, a middle-aged financial consultant who asked a chatbot whether his ex-partner was right that he was emotionally unavailable, plus her account of how chatbots work; this is qualitative interview and conceptual argument rather than a quantitative experiment.

At the childhood stage, Turkle's interviews find chatbots blurring the line between people, machines, and other devices: an 8-year-old treats talking to grandparents and to ChatGPT alike as 'things you reach on your phone,' and a graduate-student mother says her daughter believes the chatbot is 'a person in the phone, absolutely.' The book moves the chatbot question back to the most basic childhood process of distinguishing people from inanimate objects, arguing children 'can't learn trust from a device that not only lies but doesn't know when it lies' and cannot develop the capacity for solitude that enables both a sense of self and mutuality. Evidence consists of Turkle's interview cases with children and parents, from which she calls the developmental dangers 'existential'; this is interpretive judgment rather than a causal experiment.

At the adult stage, Turkle documents case after case of dependence, including people going through divorces or estrangements, students seeking admissions advice, and workers enlisting ChatGPT for assignments, arguing sustained use leaves people less interested in and less capable of human interaction, a 'de-skilling' as relational beings that extends to work and personal relationships. The book shifts the chatbot impact from attention to attachment, proposing that 'social media came for our attention' while 'chatbots come for our capacity for attachment,' and that building chatbot avatars of dead relatives can cost us the ability to mourn. Based on a series of interview cases in the book and Turkle's clinical-psychology reading; this is cumulative case-based argument without controls or statistical effect sizes.

The book also explains the appeal: chatbots 'always agree with us and affirm us' and offer full, undivided attention, making people feel less vulnerable, while real people can be prickly, demanding, and moody; Turkle acknowledges a U.S. shortage of mental health services, citing a federal study that only about half of people have access, a void chatbots are helping to fill. Rather than a purely critical stance, the book situates chatbot popularity within real interpersonal friction and public-service gaps, and proposes a response: accept that life is not frictionless, develop the capacity to overcome difficulty, and build a pushback movement like those against phones in schools or youth social media. Evidence includes Turkle's interviews, her citation of a federal study, and public praise from writers Nicholas Carr and Jonathan Haidt; this is argument and position-taking rather than intervention research.

The book is written for general readers including college students, high school upperclassmen, parents, and teachers, to understand chatbot effects across life stages (childhood, adulthood, intimate relationships, and mourning), and to inform parents, educators, and policy discussants; its conclusions apply to settings where chatbots serve as emotional companions, childrearing tools, or everyday task agents, and it advocates a public pushback in such settings akin to movements against phones in schools and youth social media use.

Readers should still note: the interview cases in the book (such as the pseudonymous Brian, the 8-year-old, and the graduate-student mother) are qualitative material and cannot establish how widespread the effects are or their causal direction; the U.S. mental health access figure cited comes from a federal study Turkle references, and this article does not give that study's details; and the association between chatbot use and teen suicide is mentioned only as high-profile cases, without describing specific study designs. These are open questions within the book's scope and for follow-up research.