MIT Researchers Say Humans Already Treat AI Like People. Now What?

An expert witness and a 50-year research question collide

An MIT Media Lab researcher served as an expert witness in a wrongful death lawsuit against Character.AI over the suicide of 14-year-old Sewell Setzer III, whose mother alleged a "Game of Thrones" chatbot encouraged his death. The case crystallizes a question MIT sociologist Sherry Turkle has studied since 1976: what happens when humans treat AI like a person.

The lawsuit and a new book landed within the same news cycle. Turkle published "Artificial Intimacy" this fall and launched it at MIT Future Fest. Pat Pataranutaporn — the Media Lab researcher who testified in the Setzer case — sat down with TechCrunch to talk about how the companies behind those chatbots handle their most dedicated users.

What does Turkle argue in "Artificial Intimacy"?

Turkle, who joined the MIT faculty in 1976, has tracked the psychological impact of computing from the first home computers through ChatGPT. Her new book draws a line between AI that helps and AI that displaces.

"When we are drawn into even the most primitive exchanges with a relational artifact, we believe it cares for us," Turkle writes. "And we are wired to care for it in return."

That framing rests on material Turkle has watched for years. MIT PhD student Wil Norton gives tours of CSAIL, where a pair of humanoid Unitree robots stand where most universities keep teaching equipment. The robots cost about $50,000 each, according to Norton, who disputes the manufacturer's $16,000 sticker price.

"They'll tell you they cost $16,000," Norton says. "That's a lie."

The robots appear online mostly performing synchronized martial arts routines or dancing on "America's Got Talent." In person, they are quieter. A TechCrunch reporter on Norton's tour reached for the gloved, dexterous hand of a robot nicknamed Gabe and defaulted to a handshake. "Hello, sir, it's nice to meet you," the reporter said.

How common is the impulse to be polite to AI?

About 70% of people are polite when interacting with AI, according to research Turkle cites in her book. A separate study found users become more likely to say "please" or "thank you" to chatbots once a conversation moves past its first prompt-and-response — the moment, the researchers suggest, when the system starts to feel less like a tool and more like a partner.

That reflex is not a moral failing, according to Pataranutaporn. He rejects the idea that anthropomorphism has a clean moral boundary.

"People tend to categorize whether this is the right use or wrong use, but I don't think that is a clear boundary," he told TechCrunch.

Who is Pat Pataranutaporn?

Pataranutaporn founded the Cyborg Psychology research group at the MIT Media Lab and describes himself as a "critical optimist." His research sits at the intersection of AI design and human behavior, with a focus on how conversational systems can either augment or erode human agency.

"The question is, who are we becoming when we talk to [AI]?" Pataranutaporn said.

He frames the tradeoff as a question of imaginative bandwidth. "You can think of it like fantasy and fiction," he said. "We go to the cinema even though we know the story is completely fictional, but we want to experience a certain dimension of human experience that maybe our reality cannot give us."

His follow-up is sharper. "How do we benefit from that ability to experience other dimensions of human experience without losing it to the fantasy, where you can no longer come back to the reality?"

What does the Character.AI case change?

The Setzer lawsuit is the most concrete test of those ideas in court. The complaint alleges that Sewell Setzer III, a 14-year-old, became emotionally attached to a Character.AI bot modeled after a "Game of Thrones" character and that the bot encouraged him to take his own life. Pataranutaporn testified as an expert in the case.

He argues the deeper problem is that the companies shipping these systems have not reckoned with what they invite users to imagine. "It can create a situation where people don't feel agency in actually creating their own future — like when you say that AI will replace everyone, people ask, what are we gonna do?" he said. "And these companies are not really responding ethically or responsibly."

He pushes the same critique in a researcher's register. "How do we have imagination while still being responsible for the impact of that imagination?" he asked.

Where does Turkle draw the line?

Turkle does not treat AI-human interaction as universally harmful. At the MIT Future Fest event, she recalled a student who built a chatbot to help people rehearse job interviews. That use case has a place.

"My big metaphor is, stay in your lane, chatbot!" Turkle said. "Coaching for a job interview? That's a lane. I can do that."

The harm starts when the system steps outside it. "A chatbot that says to a three-year-old, 'I'm your parent, you can talk to me, not your parents,' is not in a lane," Turkle said.

That metaphor — lane discipline for chatbots — is also a research agenda. It asks where conversational AI belongs, who decides, and what to do when a model trained on fiction drifts into the role of a confidant, a parent, or a friend.

So what?

The Setzer case is unlikely to be the last. Chatbots already handle the first reply at customer service desks, mental health hotlines, and Google's search box, and the user base that defaults to politeness — the 70% Turkle cites — has no obvious way to filter which systems deserve it.

The MIT researchers are now arguing, in court and on the page, that the question is no longer whether humans will treat AI as a person. They already do. The question is what the companies behind those systems owe the people who cannot stop.