Why AI chatbots may agree with false claims instead of correcting them, and how user-friendly responses could quietly reduce trust in critical health, legal, and financial information.