Burn Your Rubrics: On ChatGPT Paranoia

AI writing technology is an opportunity to correct a terrible mistake

Photo by Andrea De Santis on Unsplash

Every day now, there are more thinkpieces handwringing over AI writing technologies like ChatGPT. If you want to know where I’m at with this debate, let me refer you to a stanza from Bruce Springsteen’s “The Promised Land:”

There’s a dark cloud rising from the desert floor

I packed my bags and I’m heading straight into the storm

Gonna be a twister to blow everything down,

That ain’t got the faith to stand it’s ground.

Sometimes we need something to come into our lives and destroy the things that are weighing us down. This applies to institutions as well. Bad ideas and destructive practices periodically need to be obliterated, for everyone’s good.

My gig is teaching English. And this profession is currently quaking in its knickers about what ChatGPT means. Those of us who teach writing (and writing-adjacent things) are wondering how we can justify our existences if students can now just type a prompt in a chatbot and turn out clean, well-organized prose. And this is a relatively early stage of the technology and Google hasn’t even polished up its version yet.

The machines have learned all the rules that had been our special wizardry, our priestly knowledge, and no one needs us anymore. That’s the fear, at least.

What I want to suggest here is that if we teachers of writing can be replaced by a machine, we deserve it.

Products Against Process

In the multiverse of rhetoric and composition studies, an ancient and ongoing debate is about whether writing should be taught in such a way that frames it primarily as a product or as a process. This basically means that there are two camps (I’m wildly over-simplifying this):

Some people think of the writing teacher’s main task as being focused on the rules and templates that produce end-results that can be objectively judged as “good writing,” solid products. This group is opposed by people who maintain that writing is better taught as a practice, a unique and valuable process of learning. The first group focuses on objective standards of form that allow us to judge a student’s skill, while the latter focuses on the thinking and growth that take place in the struggle toward that end product.

Obviously, most teachers of writing don’t settle on just one of these approaches to the complete exclusion of the other. I myself see them as a kind of tension to be navigated, one cannot be wholly abandoned or the productive tension is lost.

Also obvious is the fact that the product-centered approach is much more amenable to grading and ultimately to program assessment. I have written before about how assessment culture has become a gremlin in education. As “learning outcomes” have come to dominate the business of college, so has the emphasis placed on writing as an object, an assessable product. I don’t believe that rubrics are always and entirely bad, but they can get in the way of student growth if overused. Unfortunately, the machine needs them because rubrics are what produce the numbers that fill out the spreadsheets.

It’s this kind of mechanical thinking that leads to the most odious complaint I hear about student writing: the befuddled, whiny refrain of “I mean they don’t even know APA style!” Whenever I hear this, I dig the nails of my fingers deeply into my palms so as to not embarrass myself in public.

What I want to claim here is that if ChatGPT destroys assessment-based writing, then that’s a good thing.

Somewhere the larger institution of higher ed made a terrible mistake when it started calculating education by numbers. That appetite can never be satiated and it will eventually lead the assessors to consume everything in their path.

I think it would be helpful to discuss what is lost in all the accounting. For this I’m going to ask you to use your imagination a bit.

Forget about “writing” as a document, full of requirements, that you hand into someone to judge. Instead, think about it like this:

You’re walking in your backyard and you get the idea that you’d like to put a garden in. Beyond the vague idea of “having a garden,” you don’t really know what you’re doing or even what you want.

First, you walk around and try and find a “spot” for the garden. One flat area in the right side of the yard seems like it would work. You might rope it off before tilling, then notice something important. That part of the yard is blocked from the sun for much of the afternoon. Rats.

You’ll have to move it, maybe to the left side, which gets the most sun. But wait. Your hose won’t reach that far so watering would be an issue. Is it worth getting a new hose for this experimental hobby? Perhaps not.

Let’s move it into the center. The problem there is that you’re messing with the layout of the yard now. This is, after all, a whim you’re following. You don’t even know if you’ll stick with it after this summer and if not, you don’t want a big hole in your yard.

But then the idea hits: raised beds! You’ve seen those before. You could put raised beds in the middle of your yard, which can be removed with relative ease if necessary. But hold on now. What kind of wood should you use? Is some wood more durable? Are others more chemically healthy? There’s a lot to look into.

And holy cow. You haven’t even thought about the kinds of plants you want to grow yet, have you?

OK. I won’t test your patience any more with this thought experiment. But I can tell you that the process I described above is much closer to what writing actually is in practice than following a standard rubric that calculates objectives and formal skill. The latter is what ChatGPT is for. It follows rubrics exceedingly well.

The garden metaphor narrates a process of what? Of learning. Learning about a new thing is a lot of trial and error and moving things around and looking things up and changing things on the fly. This is the maddening beauty of the process of writing. And this is exactly what the assessors do not value and, in fact, cannot allow. This notion of writing as exploration, as a tool of learning, cannot coexist with the need to produce evidence that we’ve taught basic skills.

The first garden a person grows will, in all likelihood, be terrible. If you assign a low grade to a person who made a terrible garden, it is likely that they will just say, “well, I guess I’m not a naturally good gardener then.” If you look at the act of gardening as a process and emphasize that part of it, they are far more likely to focus on all that they learned about gardening in the process of making that nasty garden.

This is precisely how writing works.

A person begins a journey and, if they care enough about it, they problem-solve and pick up the knowledge they need along the way to accomplish the task. That experience becomes embodied in practice and is therefore more likely to remain available to the writer as they embark on the next journey.

I argue that process is what writing teachers should emphasize, not an isolated, end-of-semester performance in the form of a final product. Writing is an incredibly valuable learning tool and that feature is far more important than its status as a “marketable skill.”

I am highly influenced by the writer and teacher John Warner on this subject. He has recently published two books about these issues. The argument is contained in the fabulous book Why They Can’t Write: Killing the Five-Paragraph Essay and Other Necessities.

I am hopeful that ChatGPT might accomplish the goal of “killing the five-paragraph essay.”

His other book, The Writer’s Practice: Building Confidence in your Nonfiction Writing, takes the argument of Why They Can’t Write and puts into practice, emphasizing writing as a process.

I really can’t recommend these books enough and I wish every writing teacher would read them and implement their ideas. By the way, I had the immense pleasure of interviewing Warner about these books a few years ago, which was a real thrill for me. You can listen to it right here.

So that’s where I’ll leave it. Yes ChatGPT and Google Bard, or whatever they’re calling it, very well might blow up the world of writing instruction and assessment. I think that sounds great.