The Misinformation Machine: Why I Don't Use Generative Artificial Intelligence
It seems like practically every other author out there is posting something about how much they do (or don’t) use generative artificial intelligence (AI) in their writing. So I want to let you know—I don’t use generative AI at all. Period. Everything you see in my blog posts, articles, or books is 100 percent written by me (or properly cited if it's a quote from someone else). Every single letter was typed by my fingers on my very worn keyboard.
Since I started working as a copyeditor, I’ve had some friendly disagreements with my editor, Paul, about how much AI tools can help with editing. He’s more of a techno-optimist than I am. I generally take a wait-and-see approach to new technology—let other people work out the initial bugs and glitches, then see if the improved product would save enough time to make it worth buying. And so far, I have absolutely no desire to use generative AI’s writing and research tools.
I’m not impressed with the quality of AI-generated (or heavily assisted) writing. They call it “slop” for a reason. It’s paragraphs and paragraphs of grammatically correct sentences that say nothing. Paul had me try editing an AI-generated book once to make it “sound like a human wrote it,” and after a couple pages I realized that the only way to do that was to totally re-write it.
But it’s not the slop that scares me about AI. I’m not worried about computers taking over the world, either. No, the AI application that I personally find most disturbing is when people use it as a research assistant. Even journalists are doing that now, relying on an AI-generated summary for facts instead of looking at primary sources. The big, huge problem with that is that AI always sounds like it’s right, but almost half the time it isn’t, especially for obscure or controversial topics. The only way to know for sure if it’s accurate or not is to fact-check. And by the time you do that, you might as well just do the research from scratch—which is what I do instead.
I recently edited an article that I could tell right away had been written with AI assistance because it had en-dashes, which nobody but copyeditors know about because they aren’t on a normal keyboard. (I do use en-dashes myself now, but I insert them manually). I gave the author the benefit of the doubt and assumed that she just used AI to clean up typos—until I started checking the perfectly formatted citations and found that 10–15 percent of them didn’t exist. I call them “phantom citations.” The technical term is “hallucinations.” And they made me question the whole article. If the author didn’t check the citations, did she fact-check the AI-generated content? And if the citations weren’t real, why should I trust anything the article said?
This raises a troubling question—why did the tech companies develop tools that are based on plagiarism and generating inaccurate information? This isn’t a glitch in the program; it’s baked into what generative AI is. We’d never tolerate this kind of inaccuracy with math. Imagine if a company made a calculator that sometimes did the math right and sometimes threw in some random numbers that messed everything up: “Let’s see, this pack of gum costs $1.29, and you gave me a $20 bill. Let me plug this into my calculator—okay, your change is $5.50.” So why does anyone even consider using such an unreliable tool for writing or research?
It’s a sad reflection on our society that we’re pouring vast amounts of energy and resources into a misinformation machine—and then telling people they can listen to it instead of thinking. That’s not progress. That’s just sick. Artificial intelligence is like artificial sweeteners—it may fool you at first, but it doesn’t satisfy, and it leaves a bitter aftertaste.
Comments