ARTIFICIAL INTIMACY: Who We Become When We Talk to Machines, by Sherry Turkle
Beware the robots. First, they seduced us by relieving us of drudgery and routine tasks; now, they’re coming for our souls.
So suggests Sherry Turkle in “Artificial Intimacy: Who We Become When We Talk to Machines,” her mournful new book about the acceleration of human attempts to seek connection with chatbots.
Turkle, who was trained as a psychologist, has written several touchstone volumes about our emotional relationships with technology. In “The Second Self,” published in the ancient days of 1984, she was already exploring how interactions with computers were affecting people in profound ways. Turkle was occasionally critical, but not despairing. “Computers are not good or bad,” she concluded. “They are powerful.”
More than four decades later, computers may still not be good or bad, but Turkle has seen how we have succumbed to their power. Artificial intelligence has advanced to the point where we now confide in our machines, turning to them for guidance, reassurance and comfort. We have invited them not just into our homes but into some of the most private recesses of our emotional lives. “By 2025,” she writes, “the most common use of chatbots was for therapy and companionship.”
Turkle is an elegant writer; she uses the precise and accessible language of someone who has thought deeply and deliberately about her subject, and has followed its changes over time. In earlier books, she was already becoming wary of technologies, like social media and texting, that were nibbling away at our relationships with one another. She has long argued that, to our detriment, we prefer the control machines seem to provide to the unpredictability and vulnerability of face-to-face conversation. But in “Artificial Intimacy,” Turkle seems to have abandoned wariness and skepticism for full-on alarm. Her first chapter is titled “State of Emergency.”
The book is largely organized around what she calls “the arc of life,” including chapters on “Childhood,” “Adolescence,” “Work” and “Mourning.” At every turn, she notices how machines have been given roles that used to belong to humans. ChatGPT writes a high school student’s paper and talks knowledgeably (and indefatigably) about Thomas the Tank Engine to a 4-year-old boy.
It gives a woman named Naomi the kind of attentive, “on demand” therapist she says she has always wanted. She recounts how, during a particularly infuriating conversation with her father, ChatGPT helped her calm down by guiding her through some breath work. “I always feel better when I talk to Chat,” Naomi says. “I feel a lightness.”
This enthusiasm turns out to be a theme. Most of the people Turkle talks to either heap praise on chatbots or are resigned to their ubiquity. The enthusiasts are more disturbing than the fatalists. A mother named Norah is happy that her 10-year-old daughter “vents” to Alexa: “When she talks to a person, she is very shy, and tries to be a good girl. But when she talks to Alexa, she is so open, so engaged. It’s fantastic.”
Turkle understands Norah’s relief, but suggests that Alexa is preventing the daughter from learning a crucial lesson. Norah buys a moment of tranquillity while her child never actually learns how to express her anger in the context of a human relationship.
“People become real to us by frustrating us,” the psychoanalyst Adam Phillips once wrote. “If they don’t frustrate us they are merely figures of fantasy.” A chatbot gives us what we want — or what we think we want. It isn’t programmed to challenge us the way humans invariably do. It aims to please. Any impression that a chatbot may be looking out for our interests is ultimately just a bid for continued engagement.
“Artificial Intimacy” inevitably includes the awful stories of people who have killed themselves after talking to a chatbot. Sometimes the chatbot offered an “anti-suicide pitch,” as Turkle puts it, before changing the subject. But in one especially horrific case, a 16-year-old boy asked ChatGPT for advice on methods he could use to take his own life, and it supplied the requested information. ChatGPT also discouraged him from letting his parents know about his plans, instructing him to “make this space the first place where someone actually sees you.” Within a couple of weeks, the boy died by suicide.
Turkle contends that people turn to machines out of loneliness, which leads to more loneliness. Chatbots are the snake oil that sickens us further while pretending to be the cure. The book repeatedly shows how they encourage us to abandon the genuine difficulties of meatspace for the phony consolations of technology. We may feel momentarily better while experiencing more sustained isolation.
She writes about a time she saw a nursing home resident, whose son had cut off contact, talking to a robot shaped like a baby seal. The robot would make comforting noises when it sensed sadness in the resident’s voice. But the robot had no experiences of family or abandonment or the fear of dying. “I was shaken,” Turkle recalls. “We had come to expect more from technology and less from each other.”
She is generally dismissive of the idea that robot companions are better than nothing, but there have been fascinating and complex stories about such devices helping elderly people manage loneliness, as well as their desire to continue living in their own homes. I wish the book had dealt more substantively with this conundrum. A teenager sucked into an A.I. relationship is one thing; an octogenarian finding comfort with a soothing robot doesn’t seem so obviously harmful to me.
“Artificial Intimacy” doesn’t so much build an argument, with the mounting drama that entails, as circle around Turkle’s abiding preoccupations. She writes with such clarity that it doesn’t take long to grasp her point. For anyone paying attention to the news about A.I., this book isn’t filled with many surprises. Turkle is worried about how our machines are changing us, but she is also worried about how our machine-coddled selves are retreating from the world.
Still, I wasn’t entirely prepared for her observation that our increasing intimacy with machines is destroying our democracy. In her closing pages, she warns that chatbots have been distracting and pacifying us when we need to be ever more engaged in civic life. A.I. has encouraged our descent into “passivity and post-truth,” she says.
The assertion rings terribly true. The world feels stressful, so we turn to our tools for artificial intimacy, which make us so “accustomed to hearing pleasing nonsense” that we can no longer handle unpleasant reality. “When people talk about how relaxing it is to talk to a chatbot,” Turkle writes, “it’s because the real slips away.”
ARTIFICIAL INTIMACY: Who We Become When We Talk to Machines | By Sherry Turkle | Little, Brown | 262 pp. | $32
The post Why Confiding Your Woes to a Robot Isn’t Going to Help appeared first on New York Times.
