To AI or Not to AI

I recently wrote about how some people—including writers who say they oppose AI-generated writing—get defensive when questioned about their own reposting of AI content.

I decided to post that Substack on my personal Facebook page. Because I know Facebook is allergic to links, I copied the full text into the post. I know that’s a lot words, but it was the only way to avoid the link.

More than one person commented that they don’t normally read long Facebook posts anymore. One said she only read mine because she figured I’m someone who wouldn’t use AI.

Basically, what’s happening is that people are associating long posts with AI. That hadn’t occurred to me before, though maybe it should have. But now that I think about it, I have been seeing tons of long AI-generated social media posts. So I guess, more specifically, what didn’t occur to me was how that use of AI might be affecting reading choices.

Let’s consider the implications.

On the positive side, I appreciate the possibility that people might refuse to read AI. Nothing would stop AI-generated writing in its tracks faster than people ignoring it. Of course, that hasn’t happened to any noticeable extent yet.

Case in point, I just saw a writer’s post discussing some experiment she had done using AI to revise her previously published human-written book. But I got one sentence into her post—which was long—and saw that it was written by AI. So I stopped reading. But many people were reading and commenting, so presumably it didn’t bother them (if they even noticed the AI).

On the other hand, it’s scary to think people might not read a long post I write simply because they assume it’s AI. Because AI is so pervasive, they don’t stop to ask, “Is it AI or not AI?” They just see long and turn away.

This means if you have something substantial to say, and you don’t want people to pass over it, you’re left with no leeway to elaborate. You’re limited to what’s essentially a tweet. Or you have to stay within designated social media zones (e.g., Substack) where people still tolerate full exposition.

We’re already living in a world where nuance is lacking, where information about serious subjects is reduced to sound bites and people snap at each other with one-liners. To my mind, we need fuller, deeper communication, not greater brevity.

It’s ironic that AI is so long-winded. Have you noticed how long, for example, a ChatGPT answer to a simple question is? I suppose if it doesn’t take longer to write a 1000-word answer than a 100-word answer, that makes sense. Whoever trained chatbots must have figured humans would be impressed by wordiness. Even if it forces us to skim.

When I said to AI or not to AI, I was being somewhat facetious. The answer is obvious: We don’t want to be AI. But Shakespeare’s was an existential question, and AI is an existential question of at least that magnitude for us now.