The Hidden Cost of Using AI

By Michelle Barker

As authors, we already know about the ethics of using AI. How it’s plagiarizing our work and not giving us any credit, spitting out generic novels that people are uploading to Amazon hoping to get rich quick (good luck with that). How beta readers are feeding our work into a machine and sending back a report (for a fee) that is basically useless.

As editors, we cringe at the idea that people are using AI as their main source of feedback. To be clear: AI can correct your grammar, there’s no denying that. But if you want creative and thoughtful solutions to your prose dilemmas and structural woes and you think AI will take the place of a human editor, you’re in for a disappointment.

For many of you, I may be preaching to the choir, but I think there are more people out there who rely on AI than they’re willing to admit. It’s a little like fessing up to watching porn. We claim to be high-minded, but in the privacy of our offices we’re asking Chat to help us craft a paragraph or plan a scene or write a blog post.

If you think you’re fooling anyone, guess what? You’re not. Anyone familiar with AI will recognize its unique (and annoying) speech constructions and rhythms, its weird confidence and superlatives, its overuse of em-dashes (my God).

Here’s the truth (no hedging). (Hint: that’s an AI-ism.) AI is not just flattening out the language. It’s making us all sound the same. (That’s another AI-ism.)

But hey, let’s give AI a little credit.

When AI Is Helpful

AI has its uses, to be sure. When I was planning my research trip to Berlin and had to fit fifteen walks and museums into eight days, AI was an efficient tool for organizing my itinerary so that I wasn’t crisscrossing the city unnecessarily. It even sorted my museum visits into which ones were closed on which days.

AI helped my colleague fix a broken pipe. It has helped me with recipe replacements and marathon training, and according to a recent Stanford study, it is very good at diagnosing illnesses.

AI has also stepped into the therapy ring and apparently can be effective for people suffering from anxiety, depression, and eating disorders. Therapy is expensive; not everyone who needs it can afford it, and there’s something to be said for the anonymity of a computer screen. As we all know, we tend to say things on the screen that we would never say out loud to another human. The results can be devastating (witness any number of Instagram comments), but the process can also be very freeing and allow for an honesty and authenticity that wouldn’t happen otherwise.

But There Are Limits…

I asked AI to explain my German train tickets which I swear were written in code. Had I listened to its advice, I would have missed all my trains. I asked it for walking directions from the Frankfurt airport to my hotel so that I didn’t have to spend money on a taxi. Turned out the “landmarks” it told me to look out for didn’t exist, which I only found out when I arrived in Frankfurt and wandered around in circles dragging my suitcase behind me like a clown and looking for the tall tower made of glass.

I also, at one point, consulted someone who taught history in Germany because I had questions about the emigration process. He gave me incredibly detailed information that I was super excited about until I started checking into it. Not incredibly detailed. Weirdly detailed. Chat-inspired detailed. Shockingly (or not), it was all made up. And he’s not the only one who has done that to me.

I have asked it questions that I more or less know the answer to and have been shocked by the things it just straight-out makes up with this voice of confidence that makes you think you can trust it… but you can’t.

The Hidden Cost

I’m not going to say I don’t use AI. The reason I can spot AI-speak is because I’ve consulted AI enough to know what it sounds like. But it is not a creative thinker. Sometimes its advice is worse than useless: ChatGPT helped a teen write his suicide note, after which he took his own life.

But, but… It’s saving us time, and that’s great, isn’t it? Yes. But there’s a cost: if we keep relying on a machine to do our thinking for us, pretty soon we’ll forget how to think for ourselves.

You don’t believe it will happen? It’s already happening. MIT ran a study with three groups of people who were supposed to write a series of essays while being hooked up to an EEG to record brain activity. One group was allowed to use ChatGPT, one could access Google, and one had to think on their own. The results shouldn’t surprise anyone, but they are worrisome. Not only did the group that relied on ChatGPT have lower brain engagement while working on their essays, but this engagement also progressively decreased over the time of the study. Worse, when the participants were asked to reproduce the first essays they’d written, this time without any assistance, the group who’d used ChatGPT had serious trouble remembering what they’d written. Why? Because they hadn’t really written it. They hadn’t taken the time to think it through.

Let’s not fool ourselves. We’ve been on this slippery slope for a while now. How many people in this world know how to get around without Google Maps? When was the last time you paid for something with cash, and the cashier had to figure out how much change to give you? How long did it take them?

How many people can even manage without their phones? My colleague recently went on a tech-free vacation for three weeks (the horror). Three weeks, no tech. Sounds like boot camp or, you know, prison. I admire it hugely. I want to do it. I think it’s incredible. But could I do it? Not so sure.

In Conclusion

We need to start looking down the road at where we’re going. Really looking and thinking about what we’re doing to ourselves. Our brains are incredibly powerful tools, but they’re like muscles: they only stay strong when we keep using them. The more we check out and let a machine do our thinking for us, the weaker our thinking and creative muscles become.

It’s easier to ask ChatGPT for answers. It’s faster. It’s more efficient. But is it the right thing to do? Is it better?

The author Michael Easter has a great Substack called Two/Percent, named for the two percent of people who, when given the choice at an airport between an escalator and the stairs, opt for the stairs.

It’s time to choose the stairs here. Our future might depend on it.

Michelle Barker is an award-winning author and poet. She is the co-author of two craft books: Immersion and Emotion: The Two Pillars of Storytelling and Story Skeleton: The Classics. Her fiction, non-fiction, and poetry have appeared in literary reviews worldwide. She has published three YA novels (one fantasy and two historical fiction), a historical picture book, and a chapbook of poetry. Michelle holds a BA in English literature (UBC) and an MFA in creative writing (UBC). Many of the writers she’s worked with have gone on to win publishing contracts and honours for their work. Michelle lives and writes in Vancouver, Canada.