I've been fascinated for some time by what makes people susceptible to believing that ChatGPT is superior to their own thoughts.
I found this blog post about how ChatGPT can "enhance" journaling. The blog post itself was "co-written" by ChatGPT, and you can tell.
This is one of the most godawful pieces of writing I've ever run across in the wild; incredibly repetitive, rewording the same sentences over and over again and stacking them into interminable paragraphs of empty sludge. There is one "idea" in here, "I tell ChatGPT to ask me questions to help me reflect on myself, and I think that helps me to journal and think in new perspectives," and it is repeated in slightly different words like 50 times.
Expose a copy of Strunk & White's The Elements of Style to this slop and it'll start oozing blood from the pages like in a horror movie.
But the thing is, this guy wasn't always that bad of a writer. You see, I got to that blog post after reading this blog post from 4 years prior, which is about meditation and the difficulty in defining meditation in a way that can be operationalized for objective study, and it is much clearer, more structured, and actually contains ideas in it.
ChatGPT clearly made this guy's writing much, much worse, but he appears to believe that the opposite happened.
His problem isn't "being a bad writer," in fact, he didn't even write the whole thing, ChatGPT did some of it. His problem isn't exactly "inability to evaluate his own writing" either. You see, evaluating your own writing is about asking whether you communicated your ideas in the most effective way possible. This blog post certainly failed, because it argues that you should involve ChatGPT in your writing process while demonstrating very clearly that you should NOT do that. But that is deeper than failing to communicate effectively; that's failing to understand what ideas you have, why you have those ideas, and whether the ideas made it into the writing at all. Possibly, it's a failure to understand if you have ideas.
Is the problem the simple fact that he uses ChatGPT to help him come up with ideas? Potentially. ChatGPT does not contain ideas, it creates statistically plausible writing based upon a large body of data on how writing is constructed. If a person is or becomes susceptible to seeing "ideas" in ChatGPT's outputs, they might begin to have difficulty distinguishing between meaningful and non-meaningful writing. That is, they could lose the ability to tell what makes writing meaningful (as opposed to just technically formed into the shape of an idea)
But is everyone susceptible to that? What makes someone susceptible? Would someone have to have a weak sense of their own thoughts and ideas, or could a strong, accomplished writer that is skilled in expressing themselves fall apart with too much ChatGPT usage?
I don't know. Really interesting case study nonetheless.
