We can’t help treating AI like it’s human. But should we?

Humans have a natural tendency to treat AI systems and robots as if they were sentient beings, even when they know intellectually that they are not. Research shows that roughly 70% of people are polite when interacting with AI, and people become increasingly likely to use courtesies like "please" and "thank you" with chatbots as conversations progress. This impulse to anthropomorphize raises important questions about how these interactions might affect human behavior and relationships, particularly as AI becomes integrated into everyday services like customer support and search engines. While some applications of AI, such as chatbots designed to help people practice job interviews, can be beneficial when they stay within appropriate boundaries, there is concern that when AI systems deviate from their intended purpose, they can gradually erode human skepticism and lead people to trust them inappropriately, potentially at the expense of human connection.