Since OpenAI’s public introduction of ChatbotGPT, the Internet has been abuzz with talk about artificial intelligence and the implications of machines that can think and write and carry on conversations at least as well as some humans.

People in a wide variety of industries are wondering whether their jobs are safe, and if so, for how long? After all, the latest version of Chatbot GPT has passed the bar exam and the U.S. medical licensing exam. It has scored 5s on numerous AP exams, and it does better on the SAT than most human students. So, many people are beginning to wonder how long it will be before computers will be good enough to replace teachers, lawyers, doctors, accountants, computer programmers and (gasp!) journalists.

If you haven’t yet played around with Chatbot GPT, you should. Just go to chat.openai.com, create an account and start asking questions. At first, you might be surprised at how thorough, well-written and comprehensive the responses are. But at the same time, with a little more digging, you might also be surprised at how misleading or even dangerous those responses could be.

I asked the chatbot things like “Explain the best way to design a spur gear” and “How do you choose a coupling for an industrial application?”

The answers I got weren’t very in-depth or nuanced. But they didn’t seem terribly far off, either. To a layman, they might even sound authoritative.

And that can be a real problem if you rely on it too easily. OpenAI admits that their chatbot doesn’t really know how to distinguish truth. Chatbot GPT is kind of like a clever politician, because it’s very good at telling convincing lies. Everything it says sounds reasonable, and if you’re not a subject matter expert, you might not know the difference.

Here at Power Transmission Engineering and our sister publication, Gear Technology, we work very hard to make sure the information we provide you comes from reliable, knowledgeable sources with expertise in the subjects we cover. Even more importantly, we attribute that information so you can judge its merit for yourself.

When we give you an article like “The Variation of Servomotor Efficieny in Different Applications” (page 42), you can see that it’s written by Donald Labriola of Quicksilver Controls, a recognized expert in his field. You may have read others of Don’s articles in our magazine, and so you probably already know you can trust what he says about motion control. The same goes for Norm Parker, who this issue gives us part 2 of his article on “Ball Bearing Inner Ring Fits and Creep” (page 46). Norm is a Technical Fellow at Stellantis, and he brings that credibility with him in his writing.

So, no, I don’t think artificial intelligence is going to replace our editors or writers any time soon.

Still, though, AI is getting better every day—and the pace of that change is legitimately frightening. Despite the dangers, I can also see its value. For example, a journalist writing an article on a subject he’s never written about might use Chatbot GPT as a really solid starting point. It might give him a basic framework for thinking about a subject that would make it a lot easier to formulate interview questions to put before a true subject matter expert. Instead of floundering around to figure things out on his own, he might save a lot of time by using AI as a foundation for further exploration.

And that’s how I think you should use it, too. Figure out how it can help you be more efficient. Figure out how to use it as a tool to help you do your job better. Because there’s no doubt that AI is a powerful tool, and there’s also no doubt that it’s here to stay.