Why Do Writers Use AI?
I read a weird thing this weekend. Stuff like this didn’t happen before ChatGPT
I read a weird thing this weekend. God, what a world we live in, you know? Stuff like this didn’t happen before ChatGPT and OpenAI came along.
A major newspaper ran a fifty six page summer insert for the long weekend. Fifty six pages of summer activities, photography tips, gardening tips and more. One of the pieces was a summer reading list. Books that are a ‘must read’ this summer.
The list was fifteen books, but ten didn’t exist. Like Isabel Allende’s Tidewater Dreams, and Min Jin Lee’s Nightshade Market. The authors exist. The books do not.
Once people realized the books didn’t exist, they started digging into the other articles. One article on hammocks quotes a professor of leisure studies at the University of Colorado — she doesn’t exist. Most of the articles were that way.
Some of the “professionals” cited didn’t exist. Others existed but the quotes attributed to them were made up. They didn’t say the stuff the articles said they did.
The entire article was a ChatGPT hallucination.
All of the articles had one writer byline. So Slate picked up the phone and called him.
He said he blew it and that’s on him. He’s a 56-year-old media lifer with two writing degrees using AI to try maintain an impossible human workload of low-paid gigs.
Reminded me of a piece I wrote about Keanu saying corporations don’t give a f__k about paying artists. The writer basically said it’s impossible to write that many stories for what they’re willing to pay. That’s paraphrasing, not a direct quote.
Here’s the mistake he made. I think he should have indicated that he used AI.
According to the Library Of Congress Copyright Office, machine generated text cannot be copyrighted to a human. So use of AI should have been cited. But I doubt if they’d have paid him if he put “written by ChatGPT” on the articles.
Basically, they want good writing, but they don’t want to pay for it. Which is pretty much what Keanu said three months after ChatGPT launched a couple years ago.
Here’s what one of my readers told me. I think I’ve told you this one before. He said he uses ChatGPT because he has so many stories he wants to tell but he doesn’t have a decade to learn to be a good writer because he spent his life working some job that paid the bills and now he wants to share his stories. ChatGPT helps him do that.
I’ll be honest. I don’t understand that line of thinking. I don’t. Because there is no AI who can tell my stories. Can they create content? Absolutely. In my words? No.
Here’s what another writer told me. He said he’s trained a mini-ai on his own writing and now it’s pretty good at producing articles that sounds like him. Just needs to edit a bit and it’s much faster to create content. Because he’s not trying to “be a writer” per se, he creates content that helps him sell products.
Which, okay. I understand that on some level.
But to me, it begs the question — should they disclose?
According to the copyright office, the answer is yes.
But I get the reasons AI writers don’t disclose. The newspaper writer, for example. Would he have gotten paid? Does a major newspaper want articles that said written with ChatGPT? Probably not. The guy with the personal stories, would people respond the same if his deeply personal stories said written by ChatGPT? And the guy selling courses, would people sign up for his paid programs if his articles said written by AI?
I’m going to guess that the answer to all the above is a resounding no.
So it begs the question.
If you’re using a tool you have to pretend you’re not using — why use it?
That’s not a judgement, it’s a sincere question. In the case of the newspaper, I would turn down the job. Sorry you aren’t offering enough. I’d wait tables first before I’d pretend I wrote something ChatGPT wrote. But that’s me. And I’m really curious about the motivation of people who use AI and don’t disclose. What do you think?
“I'm not upset that you lied to me,
I'm upset that from now on I can't believe you”
—Friedrich Nietzsche
Funny enough, what I got from this article was an understanding that it isn’t really about the individual declaring that they used this tool but our outdated systems.
If people cannot tell your work from your AI assisted work, then there isn’t a practical difference. People are so quick to blame all the errors on AI nowadays but I actually write buggy code all on my own.
AI is a tool in the early stages. People are using it because there is this survival pressure from our system that doesn’t really reward quality as much as it rewards the illusion of work.
Full judgement from me: if you use AI to generate content, you don't deserve a penny for it, nor any credit. Full stop. AI is for the lazy or unimaginative - at best. At worst, it's borderline criminal.