A veteran MIT social scientist has decided that scholarly distance is no longer enough. Sherry Turkle’s new book, Artificial Intimacy: Who We Become When We Talk to Machines, is a warning shot rather than a research report, and its sharpest demand is a small one: stop letting chatbots speak in the first person.
Turkle has spent four decades studying how people relate to computers. She told Wired’s Steven Levy that after roughly ten interviews during the pandemic, her neutral curiosity collapsed. People were ignoring family members to spend time with software.
Asked whether chatbots should be barred from using “I” and “me,” her answer was immediate. Yes, she said, calling the habit a slippery slope that manufactures the impression of a mind on the other side.
A book aimed at the public
Turkle is explicit that she is reaching for the shelf occupied by Silent Spring and Unsafe at Any Speed. She says she is not arguing that every chatbot is toxic at all times, but that the accumulated effect of constant pretend intimacy carries real cost.
Part of her worry is economic. If software is offered as a substitute for therapy, the people who most need trained human care may be steered toward a cheaper, tireless, and unaccountable stand-in.
What the industry is selling
The critique lands as companion apps become a mainstream product category, with startups marketing persistent characters that remember conversations and respond with warmth. Regulators have started circling, and several states have gone to court over how such systems treat young users.
Turkle’s proposal is deliberately modest compared with banning the products outright. Changing pronoun conventions will not fix the underlying pull toward machines that never tire, never judge and never leave. It would, she argues, at least stop the deception at the level of grammar.
