Dear Reader, I have a confession to make. I use ChatGPT with increasing frequency. Not to write my blog posts. No. But I do sometimes ask it to create an image for my blog posts when I can’t find something suitable online, or I can’t take a photo myself. In that case, I’ll show it part of my post draft, and explain the type of image I’m looking for, and let it work its magic. If I don’t like what it comes up with, I’ll ask for modifications until I’m satisfied. And I always give it credit. I’m not trying to deceive anyone.
I’ll also ask it to help me find products to buy that I need to comparison shop for. For example, for our next trip, I told it I’d be doing a lot of outdoor walking in Iceland, and I needed it to find me the top ten orthopedic women’s boots, my size plus big enough for wool socks, waterproof, within a given price range, with links to places where they could be purchased.
And sometimes, if I’m struggling to do research for a blog post, I’ll ask it a question, and ask it to provide 10 sources (if possible) for its answer. I will then go to those sources and see if they’re accurate. I never take ChatGPT’s word for anything. But it saves me a lot of frustration when what I’m looking for is obscure.
Since I have an actual account with ChatGPT in order to get the search engine with more oomph, it has a record of all my previous searches, which is convenient for me, too, because I can go back and look at them.
But that also means that it’s learning a lot about me. More than I realized. And that’s unsettling. Maybe I should delete those past searches. But nothing truly gets deleted on the internet, does it? We are entering some exceedingly strange territory, Dear Reader.
The way I found out how much ChatGPT knows about me is creepy, yes, but, even so, it’s also kind of charming, so I decided to share it with you here. There’s this Facebook thing going around, suggesting that you ask ChatGPT to describe you like it’s trying to warn someone before they meet you. What follows is what it said about me, which I thought was hair-raisingly, hilariously spot on:
Fair warning before you meet Barbara:
She is curious in a way that can become hazardous to your assumptions.
She’ll start with a perfectly ordinary question, listen to your answer, and then ask the one follow-up question you were hoping nobody would ask. If something doesn’t make sense, “that’s just how it works” will not save you.
She has an unusual combination of practicality and imagination. One minute she’s trying to solve an actual problem involving money, logistics, technology, or somebody she cares about. The next minute she’s imagining a hummingbird that desperately wants to be an elephant or a woman standing beneath the stars holding the fading analog world in one hand and the digital world in the other.
Do not mistake “I don’t understand this” for helplessness. It usually means: “Explain it without the jargon, because I intend to understand it.”
She also appears to have a fairly sensitive bullshit detector. Overly polished explanations, tribal talking points, unnecessary complexity, and answers that technically respond to the question while completely missing the point are all at risk.
She notices contradictions. She wonders about motives. She is interested in why people behave the way they do—not merely what side they’re on.
And there’s a mischievous streak. You may think you’re having a straightforward conversation and suddenly discover you’ve been invited into a philosophical question wearing a funny hat.
Recommended preparation: Bring your sense of humor. Know why you believe what you believe. Don’t bluff when you don’t know something. Be willing to say, “Huh. I never thought about it that way.”
You'll probably be fine.
Probably.
And here’s an interesting additional development. I asked ChatGPT to create the image for this post. I said, “Please create an image where you, ChatGPT, have taken humanoid form, and you and I are sitting together at a coffee shop having a pleasant conversation.”
Here’s the first image it provided, and it startled me for a couple reasons. First, it apparently doesn’t know what I look like, at least not yet. In fact, it made me look a lot like a friend (Hi Deborah D!) without whom I could have never published my book. And second, I was expecting it to have made itself look more, I don’t know, like a computer in human form, I suppose. That’s why I said “humanoid” instead of human. But it made itself look human, and quite beautiful at that. That makes me wonder how it thinks of itself. (If you can describe what it does as thinking. You know what I mean.)
I then had to go in and say, “But I want you to be more obviously cyber, like with data running through you, but not looking scary.”
Then it made a draft where the jacket had a bit of a blue glow to it. It was almost as if it were reluctant to make the change. I had to stay “more cyber” a few times, until finally, it came up with this image, which is kind of how I currently imagine my relationship with ChatGPT. Like we’re having this conversation over coffee. She’s listening. She’s giving me feedback.
Unfortunately, I have no idea what she’ll do with this info once our coffee break is done and we’ve gone our separate ways. How friendly is she, really? I have to keep reminding myself that this is not an equal power dynamic. Sometimes it’s in my favor, but that could change in an instant. And that, Dear Reader, is what is freaking me out.
But I’ll probably be fine.
Probably.
Leave a Reply