Comment by xnyan

I was on a technical interview panel last week where some one of the candidates was using a tool that transcribed our call and fed answers to the candidate in real time.

The first problem was the answers were extremely AI. You could instantly tell that everything they said was kind of off strange. I would not be surprised if this is something that can be improved or addressed, however what will be harder to solve is the fact that the candidate was clearly reading answers. You could see their eyes flick up and parse the text the model was feeding them very clearly. It could not be more obvious.