This year my team is facing the challenge of scaling the application we take care of, which is why I've been trying to find a new developer to join us.
As part of our process, we send a coding challenge to the candidates in advance, giving them some time to send us an answer. Based on that, we can get a picture of the technical skills available. So, when we invite candidates to an interview, we expect them to show us their solution, and together going through the code, we provide our feedback.
Recently, my colleague and I had one of such interviews. After going through the code, we continued asking a few questions about general technical topics like testing (strategies, automation, etc.) and design patterns. We were already noticing that the candidate was not too eloquent: long and a little bit unclear answers, but at the end making a point somehow. We were still seeing the candidate's screen, the "Stop Sharing" button had not been pressed yet.
So when we asked: "please tell us about the MVC pattern", we could not believe our eyes when we saw how the candidate pulled up a browser tab with ChatGPT open, and typed "mvc pattern". After mumbling a little bit (saying something like "eehmmm") while the answer was being generated, the candidate simply told us what was written there.
We were confused, but somehow managed to carry on with the interview till the end. Apart from that incident, we were not convinced anyway due to many other aspects not relevant to this story. So we did not hand in an offer in this case.
But that thing with pulling up ChatGPT in front of our eyes was so unbelievable, that I felt I needed to share this story. How can someone cheat like that? Not only cheating, but showing it openly?! Because I can't believe the candidate just forgot that the screen-sharing was on. Not in today's working world, where we constantly need to share our screens in some video call.
My personal learning: as an interviewer, I need to find a way to avoid being fooled by ChatGPT answers in the future. Somehow be more clever on the type of questions asked, and refine the way I get my impressions from candidates. There is got to be a way to still assure the 100% humanity in the face of the AI revolution :P