AI can supercharge your troubleshooting powers. It can also destroy them.

By now, you’ve probably used ChatGPT or Claude to figure out some kind of tech issue in your life. Maybe you asked it for an explanation about a weird error message you got when you tried to check your email, or it helped you narrow down what part of your website was causing a critical error, or where that funny noise was coming from in your car.

Great, right? It’s super helpful! And if you rely on it too much, it will weaken your own ability to troubleshoot.

There’s a reason good schoolteachers don’t just tell you the answers. Your brain needs to find connections itself in order to learn. Through repetition and pattern recognition, we gradually acquire new skills. This might manifest as getting better at understanding a foreign language after a year of Duolingo, baking a mighty fine loaf of sourdough after several attempts, or becoming a journeyman plumber. All of these abilities take time to develop.

Understandably, getting AI to solve your tech problems is very appealing. It lowers the stress of facing the unfamiliar and uncomfortable. Often, it gives you solutions that you can quickly implement without having to wrap your head around complicated concepts, saving you time and mental effort. If the solution worked, you can quickly get back to what you were doing.

But using AI comes with a personal cost. The next time a similar problem appears, you won’t have absorbed the cause and the solution. You won’t start to recognize categories of problems – for example, is this WordPress error because of a plugin, or because of my hosting? You won’t be able to judge accurately whether AI’s answer is even correct, because you have nothing to compare it to; it’s like you’re starting fresh each time. Over time, you become more dependent instead of more capable.

Some people will object, “Sure, but I’m learning how to use AI – isn’t that like the old complaint that if kids get to use calculators in class, they won’t learn how to do math?”

Did those kids have calculators that very often gave them wrong answers?

Other people will object, “I don’t care about getting better at troubleshooting! I just want my problems fixed!”

Fair enough. You don’t have to be an expert in everything. But when AI gets it wrong – and it does, often – how will you know? And how will you fix your problem when it’s giving you the wrong answers while assuring you with total confidence that it’s nailed the solution?

There’s evidence now that AI has already begun to impair our ability to think critically and solve problems, thanks to “cognitive offloading” or reduced mental effort. Not only that, the more confidence you put in AI instead of your own skill, the worse your critical thinking becomes. As if there weren’t enough reasons to keep a wary eye on AI, you can add “losing humanity’s capabilities and potential” to the list.

This research lines up with what I’ve noticed in my own work. AI is brilliant at speeding up the tedious parts of WordPress troubleshooting, but it’s only because I have decades of experience already that I can use it so effectively as a tool. I understand what its suggestions mean, and when it gets it wrong, I know how to articulate a correction in a way that should help it get back on track.

When I try to use it to do something in a programming environment I don’t know, though, I can get pretty lost. If the solution doesn’t work right out of the box, and I have to tell it multiple times “that still didn’t work”, solving it with AI ends up being just as slow and frustrating as if I’d tried to do it from scratch – except at the end of it, I still haven’t really learned anything.

The other problem is that I’ve seen it make enough mistakes on WordPress questions that it’s hard for me to trust it in areas where I’m not already an expert. I am hopeless at plumbing, for example. If I ask it to help me fix a faucet issue in my kitchen, can I be sure that it’s not going to waste my time – or flood my kitchen – with a cheerful but useless DIY fix?

The approach I suggest is whether you’re trying to fix your site or your ceiling fan, treat AI as an assistant, not an authority. Give yourself time to try to work out the answer on your own first, even if it’s a little outside your comfort zone. Ask it to explain concepts to you, and see if you can understand what it’s telling you to do. Don’t be afraid to challenge it – after you point out a problem, it will often admit it gave its advice about an old version of the software you’re trying to troubleshoot, or that it made up a function that doesn’t exist. 

Every time you try to solve a problem yourself, you’re building mental muscles. AI can spot you at the gym – but if you let it do all the heavy lifting, what good is it really doing you?