Me and Tom vs. ChatGPT

My good friend Tom, is an excellent expository writer. We met while working for Boys & Girls Club. Tom was hired to write grants although he had no previous experience in grant writing. We were impressed by his lengthy resume of writing and editing for newspapers and colleges and he passed our writing exam with flying colors. We felt he could learn to write grants and he did so with great success. I consider him an expert in punctuation, grammar and word choice.

To me though, Tom’s writing is a little superfluous. Doesn’t mean I don’t like his writing; in fact, I totally respect his writing abilities and greatly value his approval, editing and suggestions of my writing.

Tom and I met for lunch a couple of months ago. I brought several entries I had written for my blog – www.mytruelovelist.com – and entries on the same blog topics “written” by ChatGPT. We had been talking about ChatGPT and what it means for writers like us and others such as students and advertisers. It was pretty clear which was which so I didn’t fool him at all, but he did give me some helpful guidance on using AI programs. And he’s using AI himself but lightly, carefully, focused.

Here’s a link to an excellent article about finding your writer’s voice written by Estelle Erasmus for the ajsa (American Society of Journalists and Authors.)

My 10 minutes is up. Wishing you a wonderful writing day!

One thought on “Me and Tom vs. ChatGPT

  1. triciajenkins2001's avatar triciajenkins2001 August 17, 2024 / 3:58 pm

    NOTE: My friend Tom, who I wrote about in this blog entry, wrote this comment in an email to me. I copied and pasted it here because he wanted to share his insights with my readers.

    “The boss has me putting everything I write through ChatGPT. So I now have lots of experience comparing what I wrote to what it wants to write. Here are my conclusions:

    1. 60% of the changes are to substitute a 50-cent word for a 25-cent word. If you love words like  “finalize,” rather than “finish,”  you’ll love AI. I was taught that when you have a choice, use the simpler word because ore people will more easily understand it. AI’s belief apparently is that you appear more professional if you use large words, as long as you use them correctly. Oh, and never say there is a problem; AI changes it to a challenge almost automatically.
    2. About 10% of the changes took what I wrote and made it smoother and more easily read. On more that one occasion, I thought, “Why didn’t I think of that?”
    3. In about 29% of the changes, it seemed to be changing for the sake of change. What I wrote and what it wrote said the same thing, and, looking at it objectively, one was not appreciably better than the other.
    4. In the few remaining cases, and only a few, AI took my words and turned them into something inaccurate. When I went back and looked at how that happened, I could see how AI took what was written and reached a false conclusion about it, either because it  did not have enough context or just reasoned incorrectly.

    Those who say that your AI is as good as the prompts you use for it are absolutely correct. In our situation, the person who controls our AI has been very good at prompts. But she, like all of us, is imperfect and that means you have to take what AI spits out and closely read it to ensure that the prose still says what you want it to say. Because at the end of the day, you still have to be in control of it all.

    Oh, and one more thing: When you use the AI, make sure it lets you know (through boldface or italics or some other kind of alert) any time it has made a change. We have yet to figure this one out, and it results in my having to spend a lot of time looking at my draft and AI’s draft, word for word, to see just where it made the changes.”

    Tom

    Like

Leave a comment