Dealing with AI Slop

I came across an article this morning that talked about how many people were suddenly having to deal with “AI Slop.” Essentially, this is the result of someone using AI to get a job done, and then not meaningfully interacting with the output (or sometimes, even the input). In other words, someone gets assigned a task, and they turn it over to AI to get it done, then email whatever AI generated, dust their hands off, and declare mission accomplished.

Even as a strong proponent of AI, this is something I’ve come across, and naturally, I have a few thoughts about it.

First of all, I don’t think this is something that’s unique to AI. People have been phoning it in when it came to work for a long time. It’s hardly new to have someone not try that hard. However, what is new is how easy it is for AI to mask the small amount of effort someone put into something. Before, if you wanted to write a thorough response to someone, you wouldn’t be able to do that with a click of a button. Now, it’s often much easier to write a very long response (just have AI do it) than it is to write a short response. In 1657, mathematician and philosopher Blaise Pascal wrote the following apology in a letter: “Je n’ai fait celle-ci plus longue que parce que je n’ai pas eu le loisir de la faire plus courte.” Translation: I have made this longer than usual because I have not had time to make it shorter.

So AI has just made it so it’s incredibly easy to write an awful lot of text without actually putting much original thought into it at all. And because people use AI to do the things they’d rather not do, it’s far too tempting to just copy and paste the slop and call it done.

However, I also think that with the advent of AI, people are now less patient with poorly-written items. If something doesn’t come across as well-written or effective, it can be very tempting to chalk it up to “AI slop strikes again,” even if you don’t know for a fact it was written by AI. I’ve also talked with students who are accused of using AI to write because their papers were too good. (Suddenly the semicolon and the em dash are grounds for accusations of all number of academic evils, it appears.)

Because we all know anything might have been done with AI, we start seeing it everywhere we look, even if it’s not the case.

The solution to this isn’t not to use AI at all. (This all-or-nothing dichotomy drives me wild, but it’s where the conversation often stalls.) There’s a middle ground, where you use AI effectively to both make your job easier and also communicate effectively. This means more work, however. At times, it’s easier just to not use AI at all, but there are plenty of times when you can use it to turn a list of bullet points into a draft, and then edit that draft into something ship shape. The trick is . . . you actually have to read what you “wrote.” You also have to read what you’re responding to. And you have to think about it. And come up with your own responses and ideas. (Yes, you can brainstorm with AI, but you still need to sift through those results to find what you actually agree with.)

I tend to think that overall, we will emerge from this wild west of writing in a better place than we began. We will expect more of what we read and what we’ve written. We’ll become bigger sticklers for calling people out for poor thinking (or “slop”). True, this is very difficult to do when the slop comes from your boss, but I’m confident they’re having to deal with slop as well. At some point, we’ll all get tired of it or get fired for it, if that makes sense.

In the meantime, I’m afraid it’s going to be a bit of grit and bear it until we’re on the other side . . .