I caught myself watching an AI video on youtube a few days ago where the narrator's AI voice had deliberate instances of stammering and stuttering. It's definitely simple to have these introduce idiosyncrasies just through specific prompting.

Why would that matter? You still have to read, understand, and respond. If something is important and specific it takes longer to prompt iterations to generate my response. It's nice for spelling and grammar correction.