Researchers warn we often treat AI like a person

In short: People naturally act like AI is human, and researchers say we should be careful about how that shapes trust and behavior.

What's going on

A TechCrunch report describes a common habit. When people talk to chatbots or meet human-shaped robots, they often respond as if they are dealing with a person. In the story, a reporter meets a humanoid Unitree robot at MIT and instinctively shakes its hand and speaks politely.

MIT researcher Dr. Sherry Turkle argues this reaction is normal. In her new book, she writes that even simple interactions with a “relational artifact” can make us feel like it cares about us. A “relational artifact” is a machine designed to invite a social response, like a chatbot that talks in a friendly tone (like a customer service rep, but it is software).

Some research points in the same direction. One study cited in the report found about 70% of people are polite to AI. Another study suggests people use more “please” and “thank you” as a chat goes on.

What to watch

Researchers quoted in the piece say the bigger issue is not whether politeness is good or bad. It is what happens when people start trusting chatbots too much, especially for emotional support. Turkle’s advice is to “stay in your lane,” meaning AI can be helpful for tasks like practicing job interviews, but it should not replace real human relationships.

Source: TechCrunch AI