"Is it okay to use AI?"

At a meetup yesterday, a university student looking for a job asked a speaker whether or not it's okay to use AI for solving interview coding challenges. She feels that given how ubiquitous the use of AI in software development now, it should be the acceptable thing to do.

The answer to her question was effectively "it depends on the interviewers", and some of us in the audience also discussed the importance of problem solving as skill.

Being able to solve an interview coding challenge may get you through a coding challenge, but that's typically just a part of the entire interview process: how you work through problems, how you communicate with other engineers, and how you make trade-offs are typically assessed in engineering interviews — chances are AI won't be able to help you with those where it really counts.

What would it be like if it were okay to use AI in interview coding challenges? I suppose if it were okay for everyone to use it, then interviewers will just assign increasingly complex problems — ones that AI will likely struggle with. We'll then be effectively back to where we started: you'll still be required to demonstrate problem solving skills except that the bar is much higher now.

You should probably get accustomed to how AI is used in your field and be able to work with increasingly powerful models. But the game of keeping up with technological advancements and job security has really always been the same.