My concerns about the use of AI in the GP's surgery is summarised by one small example. As part of a health check I was asked if I vaped. I said no I didn't. The notes produced by the AI summary which they humanise by calling it Heidi reported that I "denied" vaping. Technically correct. I did say no, but where is the nuance? Denial implies accusation and suggests that I was not telling the truth. I am sure the doctor would not have used denial as the negative of choice if the notes were compiled manually because he had applied his human abilities to the answer. I have concerns. #AI #NZPol