People increasingly view chatbots as friends—but maybe not their brightest friend, new research finds

New research from Stanford University shows that Americans increasingly describe AI chatbots as they would a human friend or companion. At the same time, somewhat paradoxically, their trust in these chatbots’ competence is declining. The study raises critical questions about how chatbot interfaces should be designed to potentially discourage anthropomorphism.

Photo illustJaap Arriens/NurPhoto—Getty Images