I think it's like calculators in exams; at some point, you just accept it and make the questions harder. If you're asking questions that can be solved by ChatGPT or Google, they're probably not a good filter.
We're also reevaluating the role of the programmer at work. We're moving past the era of code monkeys who memorize search algorithms and design patterns. We're entering the phase where you give a dev the context of what the product is trying to do and the devs think of the solution that involves the drawbacks/limitations/effort of implementing this solution. Code is simply communicating with a machine, and AI is similar to a higher level language.
I think part of the problem is that it's just a language model, so it's input data is a lot of crap from the internet and a thousand listicles were people are spewing the same generic nonsense.
But personally I would not, but then I just want to get my job done as quickly and easily as possible so i can move onto something else.
We have our own implementations for certain stuff, but copilot is also accessible under certain conditions.