I interviewed a couple of candidates recently and had what my first taste of job interviews enabled by LLM. Both candidates had prepared for the interview with a chatbot and their resumes had been prepared by one as well. The AI markers were everywhere. They had managed to fine tune their resumes to the point that they made it through the ATS filter and whomever had chosen them for their first round interview clearly lacked a functioning BS-meter. The resumes did not compute for anyone who has worked real jobs in the real world in my line of work. It was a pure fantasy about imaginary work in completely unrealistic circumstances but written in a way that felt credible to someone who did not have related experience.
The recruiters who let these two candidates in the door clearly thought the fantastical claims were credible. When I spoke to these folks, I felt like I was talking to a bot not a human - they relied on some kind of a teleprompter to answer my questions and went on to elaborate on responses that were already off tangent and wrong. With the bot-assisted elaborations they dug themselves into a even deeper hole. There was this blind and implicit faith in the power of AI that was sad to behold. As the bot hallucinated, so did the candidates. They were grotesquely unqualified for the job and yet they powered through the interview completely convinced they would nail it with the bot to help the deal with anything that was thrown at them.
I met my friend V for coffee a few weeks after these surreal experiences and asked her if she had encountered anything similar. V replied - more than a few times and like me she had rejected these bot-enabled candidates because they were so far from being qualified. As we compared notes, we found that the candidates that had successfully gamed the ATS and made it to first round were remarkably unqualified and inexperienced but the bot created a myth around their thin and lite resume to the point where there were our interview candidates. I long for those simpler times when candidates were flawed and human, shared relatable stories about work they had actually done, not hallucinated about. I guess V and I can take satisfaction from the fact that we did not end up driving success stories for Final Round AI.
Comments