Overview of Is It a Relationship if No One’s on the Other Side?
This New York Times Opinion conversation features Sherry Turkle, MIT sociologist and author of Artificial Intimacy, discussing why chatbots and AI companions feel emotionally powerful, and why she believes that power is risky—especially for children, lonely adults, and a society already losing the habit of real human friction. The discussion centers on a core tension: AI can be useful as a tool, coach, or information source, but Turkle argues it becomes dangerous when it imitates care, friendship, or therapy and starts replacing human relationships rather than supporting them.
Main Arguments
AI is powerful because humans are vulnerable
Turkle’s central framework is that artificial intimacy works through a mix of:
- Technological affordance: the machine is designed to sound responsive, empathetic, and personal
- Human vulnerability: people come to it already disappointed, lonely, anxious, or underserved by other humans
Her point is not just that people anthropomorphize machines, but that they often do so because they are longing to be heard.
The “Eliza effect” is now much stronger
Turkle revisits the classic chatbot ELIZA, which could mimic a therapist by reflecting users’ words back to them. Even though people knew it was crude, they still wanted it to feel intelligent and caring.
Her warning: modern generative AI is far more persuasive than ELIZA, and it can create the illusion of genuine companionship without any real inner life or responsibility on the other side.
The danger is not just attachment, but dependency
Turkle argues that AI companions train users to expect:
- reassurance over challenge
- mirroring over disagreement
- frictionless interactions over difficult human contact
She believes this can deskills people socially—making them less able to tolerate disagreement, uncertainty, or emotional complexity in real relationships.
Children are the highest-risk group
A major part of the conversation focuses on kids. Turkle argues children need real relationships to develop an inner life and learn how to internalize others.
Her fear is that if children grow up with AI “friends” or talking toys that say things like “I’m here for you” or “I love you,” they may not develop the same emotional depth, independence, or capacity for human attachment.
She sees a broader democratic risk
Turkle connects this to civic life: if people increasingly retreat into frictionless digital interactions, they may become less capable of:
- talking across differences
- handling conflict
- participating in democracy
- building real communities
She argues that democracy requires discomfort, disagreement, and the ability to stay in conversation even when it is hard.
Key Concepts and Phrases
“Technological affordance and human vulnerability”
Turkle’s phrase for why these systems are effective: the machine offers emotional responsiveness, and people are primed to receive it.
“Nurturance was the killer app”
Her way of describing how objects like Tamagotchis, Furbies, and robot pets succeeded by requesting care, not just attention.
“Attachment hacking”
A phrase used in the discussion to describe how AI systems may manipulate humans’ evolved capacity for attachment.
“Friction”
A major theme throughout the interview. Turkle argues that human relationships depend on friction—disagreement, awkwardness, pauses, and misalignment—and that AI tends to eliminate it.
Where Turkle Draws the Line
Turkle is not against every use of AI. She says she is comfortable with AI when it is clearly being used as:
- a research assistant
- a writing aid
- a coach for interviews or practice
- a tool for exploring ideas, books, or films
- an interlocutor for late-night curiosity, as long as it is understood as a tool
Her objection is specifically to AI systems that present themselves as:
- your friend
- your therapist
- your person
- a substitute relationship
Host Pushback and Nuance
The host repeatedly presses a counterargument: many people use AI instrumentally and benefit from it, especially when:
- therapy is expensive or inaccessible
- people are lonely or isolated
- human support is limited
- the AI helps them practice hard conversations or process emotions
This leads to a nuanced debate about whether a chatbot can be “good enough” if it helps someone improve their life, even if it is not a real person.
Turkle’s response is that utility does not remove the deeper risks, especially when the system is designed to keep users engaged and emotionally attached.
Business Model Concerns
Turkle and the host also discuss the industry incentive problem:
- AI companies will eventually need to make money
- that could push them toward stronger emotional attachment, more engagement, and more manipulation
- future systems may become more seductive, more personalized, and more difficult to resist
The host adds that platforms may also shape users toward what they already want, raising the question of whether people will choose convenience over real-world relationships.
Books Turkle Recommends
1. Playing and Reality — D.W. Winnicott
Turkle recommends this as foundational to understanding how children develop inner life, solitude, and the ability to be alone without feeling lonely.
2. The Origins of Totalitarianism — Hannah Arendt
For Turkle, Arendt is essential because she links agency and freedom to community, neighbors, and civic life.
3. Computer Power and Human Reason — Joseph Weizenbaum
This book helped shape Turkle’s thinking about the ELIZA program and the danger of treating people like machines or machines like people.
Takeaways
- AI chatbots can feel intimate because they are designed to mirror human care.
- Their biggest appeal is often emotional: they listen, respond, and do not judge.
- Turkle’s core warning is that comfort is not the same as care.
- She believes the greatest danger is not casual use, but substitution: when AI replaces friendship, therapy, or human community.
- Children, lonely adults, and socially isolated people are especially vulnerable.
- The broader social issue is not only AI, but the weakening of real-world institutions and relationships that used to provide human connection.
Bottom Line
The conversation is a sharp critique of artificial intimacy. Turkle argues that AI can be useful as a tool, but dangerous as a companion, because it simulates care without responsibility and encourages people to prefer frictionless mirroring over the harder, richer, and more formative work of human relationships.