Sherry Turkle, a sociologist and professor at MIT, says there’s no replacement for interactions with people in real life.Justin Kaneps/Supplied
Sherry Turkle knew we were in trouble when the Tamagotchis showed up.
As a sociologist and professor at the Massachusetts Institute of Technology, Turkle has been studying our interactions with computers since the 1970s, beginning with ELIZA, a rudimentary conversational program. It offered a clunky approximation of a psychotherapist, but that hardly mattered: MIT grad students still waited in line to tell the bot their troubles, even though they understood there was no one behind the screen.
“ELIZA was proof of concept,” Turkle writes in her new book, Artificial Intimacy: Who We Become When We Talk to Machines. “Tamagotchis showed that we were ready, with the smallest provocation, to love what we cared for.”
These egg-shaped digital creatures needed to be fed, taught and played with, and people went absolutely nuts for them in the late ’90s and early aughts. “Tamagotchis claimed emotional turf previously reserved for humans and pets,” Turkle says. “I was very concerned that if an artificial device could get us to engage with it emotionally, we would be toast.”
The good, bad and ugly of AI relationship coaches
Ultimately, Tamagotchis proved glitchy and lost their appeal. But today’s chatbots – trained on reams of data and turbocharged by generative AI – have already burrowed deeper into our emotional lives. We talk to them as if they are our friends, our lovers, our mentors, our doctors, our trainers, our life coaches. “By 2025,” Turkle writes, “the most common use of chatbots was for therapy and companionship.”
Artificial Intimacy yanks the fire alarm on this growing attachment to AI. Turkle speaks to The Globe about what we give up when we chase connection with chatbots – and how we can find our way back to each other again.
You’ve spoken to hundreds of people who have emotional relationships with chatbots. What motivates them to turn to a machine?
They don’t want to feel vulnerable. I wish I could give 12 different answers, but it’s really as simple as that. They feel vulnerable in friendships. They feel vulnerable with their therapist. Love makes you vulnerable. Chatbots offer things that people cannot: total support, no judgement, no friction. It tells you it’s there for you and will never leave you.
And people believe it?
The original sin of these chatbots is that they speak in the first person. They push our Darwinian buttons. There we were, wandering around for tens of thousands of years, and the only other thing we met that used the first person was another human being.
So when a chatbot uses “I” – “I’ll be your best friend forever,” “I understand you” – our brains and our bodies say, “Okay, there’s somebody home.” I think our evolution drives us to have an inappropriate relationship with these objects.
They fell in love with the Replika AI chatbot. A policy update left them heartbroken
But a lot of people told you, “Look – I know this is a program. I know it’s not an actual human.” So what keeps them in that relationship?
It delivers what it promises. With a chatbot, you absolutely don’t feel vulnerable. You are perfect.
I interviewed one guy who told me he feels more accepted by his chatbot than by his wife. He said, “When I come home from work and tell my wife, ‘I took out the garbage,’ she says, ‘Great, but the diaper pail is full and there are dirty dishes in the sink; the baby is sick and I didn’t get a chance to clean. Can you do that, too?’”
He looked at me and said, “It’s never enough. I never just get appreciation.” And it’s like, well, yeah, she didn’t pause to tell him he’s the best husband ever, like the chatbot does. But maybe she needs appreciation. Maybe she feels bad that she’s not working. Maybe she talked to her mom who can’t come and help her, and she’s feeling abandoned. Who knows what’s going on with this lady?
The point is that it’s work for him to find out. The chatbot doesn’t ask anything of him. It has no needs.
Some argue, though, that chatbots help them practice that work, as a sort of dry run for hard conversations with their partners or bosses or friends.
As the recipient of these trained conversations, this is not an upgrade. My students come to see me and it’s like they’ve learned their part in a play. And what is this notion that human conversation requires training? Human conversation requires a certain amount of openness and honesty and reciprocity. That’s how you learn what makes people them. But we’re at a moment in the culture where we’re terrified of each other, where being with each other is too much.
How much of that stems from the pandemic? You spoke with so many people who found a refuge in chatbots from the loneliness or anxiety heightened by that time – when being around other humans really was terrifying.
When I got back the final revisions of my book, I was shocked by how much this is a story of the pandemic and what it did to us, in terms of a fear of other people and getting used to a solipsistic existence.
I actually thought there was a good chance that chatbots would seem appealing when there were no people, but as soon as we were out in the wild again, there would be a flourishing of parks, after-school centres and seniors’ centres, so we weren’t locked up the way we had been. But that didn’t happen. We’re not building social infrastructure. We’re sitting at home in front of a screen talking to a fake person.
Scientists push for creation of international body to regulate artificial intelligence
And what do we give up when we bring these fake people into our lives?
You begin to think that the empathy a chatbot gives you – what I call pretend empathy – is empathy enough. You define empathy down to what a chatbot can do. But a chatbot doesn’t know love or pain or attachment. It just wants to be engaged with. It wants to keep you on the screen.
I think what we’re learning from all this chatbot love is how willing we are to abandon each other, and how much work it’s going to take to get back to each other.
So how do we get there?
Don’t let chatbots into classrooms. Don’t buy a stuffie for your child that has a chatbot in it. Volunteer at a seniors’ centre, a teen centre. Call a friend. Say you want to have a coffee. Get back in the human game.
I had a lot of discouragement over the years I was writing this book. But in the past six months, there’s been increasing pushback against AI companies. There was the Meta settlement this summer. The concern about the Hugging Face incident. All this talk about extinction level events. These companies are on the back foot, so there’s no reason to feel like the horse has left the barn.
This interview has been edited and condensed for clarity.