Instead of doing useful things today, I had a status update, followed by an interview, with me frantically talking to HR about clawing back the interview they'd scheduled me for, doing the interview anyways, then some interview feedback, and then I got to spend a solid hour and a half building a document outlining for HR how to identify and bounce applicants who've clearly just used ChatGPT to prepare their applications and are wasting our time.
If you've spent some time with ChatGPT, you start to get used to how it waffles and vacillates - its answers are well-written but it refuses to take a concrete position on most things. It also loves summing up at the end of its statements. In summary, using ChatGPT to fill out your job application may or may not be a bad idea, depending on how the...
The thing is, all it takes to work around that is a little bit of prompt engineering and editing. You can simply tell ChatGPT which position to defend, give it an opinion up-front, it's not too hard to minimally conceal the provenance of ChatGPT text.
One fun hack on our side of things is that you can simply paste text into ChatGPT and ask if it thinks that it was responsible for that text. It's pretty good at detecting itself.
The ultimate lesson, though, is probably that asking candidates questions on job applications? That's probably going to have to stop being a thing.
