The overuse of ChatGPT 

Editor’s note: Alice Bronnert is a director of primary research at Optum Rx, United Health Group.

Almost every day I’m telling somebody – either a friend or a colleague – about ChatGPT. 

“Chat what?” they say. “What is that?”

Well, I’d be happy to tell you. ChatGPT is like candy. Imagine a Twizzler. It’s super fun at first. That strawberry flavor, that stretchiness, breaking it off in your mouth and then pensively chewing. It’s great for about three of these twists. And then you get to the point where you’re not so into the Twizzlers anymore. The point at which you may want to hand them over to someone else. That point of decreasing marginal utility. It’s that, “Ugh. Feeling slightly ill. Must stop,” type of feeling.

For me, ChatGPT has been a bag of Twizzlers. In a way, it’s been great. It can be used for so many things and it’s extremely fast. Here’s just a few of the ways that I’ve employed the AI chatbot:

  • Asking it to write multiple random limericks. One was about our budgie parakeets and my daughter’s water bottle.
  • Telling it to interpret the results of a very long personality and talents test.
  • Explaining a tortuously convoluted friendship situation and asking it to respond in a client-centered, humanistic counseling type of way.
  • Requesting that it explain a complicated drug pricing concept.
  • Getting it to write a fable in the manner of the King James Bible based on a squid and a watch (due to a stuffed squid and a watch hanging out on our coffee table at the same time).

The thing is, none of these interactions brought me joy (okay, maybe the squid fable a little). Mostly though, I kept asking questions because ChatGPT’s responses were fast; they were complete; they were multiple, in-depth paragraphs! But here’s the thing, the tone is unmistakably automated. There’s something off-putting about having an AI language model talk to you in a very correct, bland, pedantic and even supercilious way. I’ve often ended up annoyed. These are the times when I’ve started to argue with ChatGPT about how it needs to inject more humanity and compassion into its responses (it often retorts that it is already programmed to be pleasant).

Here’s a question for us all. How about we just stop and pause before going into this AI chat model? How about we consider something revolutionary? That we actually have the resources to do this all ourselves? We can ideate, we can create, we can analyze, we can synthesize, we can do this by ourselves. By our own carbon-based life-form selves.

For myself: I want to write and edit my own zany limericks, my own fables, my own letters. I want to be incorrect and wrong a million times. I want to be offbeat in my own weird, meandering way. I want to come to my own conclusions like Walt Whitman asked us to do when he said, “Stop this day and night with me and you shall possess the origin of all poems.”

So, I’m not feeling the love for ChatGPT. But I am feeling abiding love for myself and my fellow humans on this planet. Now if we could choose a punk, feminist, book-reading, middle-aged rocker for an AI personality, I’d be all eyes and ears. But at this point in time, I’m pretty sure ChatGPT and I need to take a break from one another. I’m saying goodbye to my Twizzlers.