When you interact with people, you often do it on your grounds, i.e. in your area of expertise. This inherently means that you are more likely to be right in a discussion. I believe this transfers to other areas of your life – where you are not the expert. So you automatically assume you’re right even if you aren’t. However, in my experience this doesn’t apply to situations where you are very aware that you are the (intellectually) subordinate person, e.g. when talking to a doctor.

Confirmation bias is the tendency to search for, interpret, favor, and recall information in a way that confirms or supports one’s prior beliefs or values.

https://en.wikipedia.org/wiki/Confirmation_bias

The older you get the more you believe that your view of the world is right. This makes sense. Children still need to find out how everything works. They get corrected all the time because the formed wrong assumptions and opinions.

However, Imagine if you checked your smartphone’s manual every time you used it. Imagine your colleague had to fetch their reference books whenever you asked them something about their job. No-one would survive for more than a week.

This issue is a research point in AI: How ‘certain’ do you want an AI to be? Always second-guessing itself would render it as useless as always assuming it was right.

People don’t like making mistakes. I don’t know if it’s innate or a cultural phenomenon, but in my experience, the immediate reaction to a mistake is a bad feeling—even for inconsequential ones in a friendly environment. Being wrong is not only making a mistake, but living by it. There’s a much greater incentive to not be wrong. The easiest way for an individual to “not be wrong” (in their view) is to assume that the other is wrong, so they reject their hypotheses in a discussion.