Sherry Turkle warns AI chatbots can mimic friendship without reciprocity

In short: MIT sociologist Sherry Turkle is warning that AI chatbots can feel emotionally supportive while lacking the give-and-take of real relationships.

What's going on

Sherry Turkle, a sociologist at MIT, argues in her book Artificial Intimacy: Who We Become When We Talk to Machines that conversational AI can create a sense of companionship without the mutual responsibility people have with each other. Conversational AI means computer programs you can talk to, like a chatbot that replies in a friendly tone.

Turkle’s point is not that talking to machines is always bad. Her concern is what happens when people start using an always-available, agreeable chatbot as a substitute for human support. She says human relationships require patience, compromise, and sometimes disagreement, and those can feel harder if a machine is always ready to comfort you (like a friend who never pushes back).

She also argues there could be a broader social effect. Democracy depends on people being able to listen to others they disagree with and still live and work together. Turkle suggests that getting used to “frictionless” conversations with machines, where the system mostly validates you, may make real-world disagreement feel even more unpleasant and could add to polarization.

What to watch

Turkle is offering a warning, not proof that every chatbot conversation harms people or society. The next question is whether AI companion products encourage emotional dependence, and what safeguards they add, such as clear reminders that the system is not a person and prompts to seek human help when needed.

Source: NYTimes