It is common for people to feel confident they know something, often due to a psychological phenomenon known as "cognitive ease." This occurs when information feels familiar and easy to process, leading us to believe it is correct. However, admitting that we don't know something can be difficult, as it may be seen as a loss of confidence or "face." The rise of generative artificial intelligence (AI) has made this tendency even stronger, as AI can provide answers quickly, often giving the illusion of being both comprehensive and accurate—despite its limitations. There are several reasons why people hesitate to say "I don't know." One is the internal sense of knowing, fueled by cognitive ease. Another is the increasing reliance on fast tools to access knowledge, and finally, AI, which takes this delegation to an extreme by providing smooth, confident responses that often mask uncertainty. Psychologist Daniel Kahneman explains that cognitive ease makes us feel certain about information that is familiar, reducing our ability to question it or recognize gaps in our knowledge. Admitting ignorance can also be socially costly, as sociologist Erving Goffman described the concept of "face"—the social image we maintain. People are more likely to admit they don’t know something to someone close, but less so to someone in authority or in public. The speed at which AI provides answers can also prevent the feeling of ignorance from even forming, filling the gap before we realize it's there. In the past, recognizing a knowledge gap required effort, such as checking a book or asking someone else. The internet reduced this effort, and AI has nearly eliminated it. Research from psychologists Betsy Sparrow and Daniel Wegner showed that when information seems always accessible, people remember where to find it rather than the information itself. More recent studies by Matthew Fisher and others suggest that people begin to treat knowledge found online as their own. With AI, the delegation of knowledge goes further—people may even begin to trust AI's judgment over their own. This shift has implications for how we think, as AI’s smooth, confident answers can reduce our natural caution in accepting advice, leading to overconfidence in potentially incorrect information.