Why I’m advising that people stop assigning essays, and it’s not just because of AI

I’m closing in on 20 sessions I’ve done in the last year that were, in some way, related to the issue of generative AI. Actually, it’s probably more than 20. They have moved from ‘omg, there’s this thing that just came out’ to ‘yeah, we really do need to sit down and have a long chat about what we’re trying to get done in higher education’. And, sometimes, that conversation is a really positive one. A humane one. We’re going to follow that tone here.

I don’t hate essays

I cut my teeth in higher education classrooms teaching academic writing. I taught the sandwich model, I did practice five paragraph essays, process writing – all the classics. I loved the challenge of getting students exciting about the writing process, teaching them to support a position, and giving them the tools they’d need to conquer the big essays that they were going to face in the rest of their university careers. When students walked into my classroom on the first day, they were confronted with this picture on the projector screen.

My position was that I had never seen a student remain neutral about the question – “Is this art?” This was the first activity in the first class.

Note: Turns out, Duchamp may not have made this. Thanks to Prof Rebecca Furguson for pointing that one out.

We have to learn to write essays because we have to learn to write essays

In the last couple of years, however, my belief in the essay has been taking some hits. I started using it as an example of an assignment type that was tethered to the system. K12 system teachers would tell me that students needed to learn to write essays so they could write essays in university. Undergrads are told they need to learn to write essays so they can write them when they are in grad school etc… I mean, what percentage of the people who learn to write an essay ever reach the point where they, actually, need to write them for some useful purpose? I mean… I did. But I’m probably the only person I grew up with who ever wrote an essay because they wanted to.

It seems like a lot of training to develop a skill that very few people are ever going to use.

But essays teach all these other skills!

So here’s the part that’s been coming up since the GenAI conversation. I’ve been using this Chronicle article to discuss how students are using GenAI to help them write essays. The author, a student, worries that students are going to lose the ability to do critical thinking because the AI is going to be doing it for them. All the student needs to do to write an essay amounts to a little grunt work.

So what are the skills that essays are meant to teach, and, if they ever worked to teach those skills, do they still in this era? I grabbed a random list from a random website

  • Analytical Skills. (GenAI is going to cover that)
  • Critical Thinking. (And this one)
  • Creativity. (And this one – I’m not saying that GAI is creative, but rather that because the choices get made the student doesn’t need to be creative to write an essay. I still think students desperately need to develop their creativity)
  • Ability to Structure Knowledge. (I mean, maybe we’re still doing this?)
  • Keen Eye for Details. (Grammarly has this one covered)
  • Informed Opinions. (Does it?)
  • Information Search Skills. (See below)
  • General Verbal Intelligence. (GenAI, Grammarly)

A quick look at this list, at least, suggests that the tools we have available are going to do a fair amount of the work that we think the essay is doing. And this doesn’t even count the fact that for somewhere between $10 and $50 a page you can just hire someone to write your paper for you. John Warner has been talking about this for years, see this post and his book.

We need all those skills, or at least most of them, but I don’t think that the essay is doing that for us anymore. I want to teach creativity, I just don’t think the essay supports that like it used to.

Search is the key. It’s all about search.

But this has been the realization for me. Essays have not been doing the thing that I actually thought they were doing since we started having effective online search tools. I used to assign essays for the same reason I used to assign writing reflections for academic papers. I want students to engage with the material. I want them to learn how to identify valid and credible information and learn how to apply it to problems they are facing. I want them to engage in the history of knowledge. With other thinkers in the field that we’re studying.

Here’s the thing. I’m starting to think it really hasn’t been happening for 20 years.

In teaching the SIFT method to my students this term, we ended up in a bunch of conversation about how we go about finding information. I heard one student say ‘i found a quote to support my argument’ and it hit me. I ask them how they’d learned to search/research, and it went something like this. Have an argument, usually given by the instructor, do a search for a quote that supports that position, pop the paper into Zotero to get the citation right, pop it into the paper. No reading for context. No real idea what the paper was even about. Search on google/library website, control +F in the paper, pop it in the essay.

Compare this, if you will, to my undergraduate experience in the library. Go to the card catalogue, write down a bunch of possible articles/books on a piece of paper, go around the library and find said resources, settle in at a table to go through them. I had to read them. I’m not saying I wanted to, there was no other option. I had to engage with the material in order to find something that I could base my paper on.

That experience is gone.

The essay, i’m arguing here, no longer forces students to learn how to research. I’m not saying a few students don’t do it, i’m saying they don’t have to. As a graduate student in one of our Humanizing Digital Learning course said to the rest of the class “You’d have to give me a reason not to CTRL F, ’cause I don’t see it”.

Teach Search

And, because of this, I now teach search. I teach students how to write good search strings to get varied responses. We explore how different searches work over different systems. We talk about bias the researchers have, about how to find facts, but also how to find advice when you don’t know what you’re doing. We talk about the humility necessary to use Internet search to learn. If you don’t have the skills to evaluate something, you’re going to struggle to get wisdom you can use from the web, or from ChatGPT or wherever you going to find it.

I’m increasingly starting to think that we need to re-evaluate what the basic epistemic skills are that we think people need to make meaning with all this abundance and all the GenAI out there. I think everyone, in every field, might want to devote some serious class time to how we can find valid and credible information when it comes to facts, but, maybe more importantly, when it comes to things that aren’t about ‘right and wrong’.

Every one of these students is going to be a voter.

I don’t think that the essay is teaching these research skills anymore, and, if anything we need them more than ever.

Further reading

I’m never Assigning an Essay Again – John Warner

Author: dave

I run this site... among other things.

16 thoughts on “Why I’m advising that people stop assigning essays, and it’s not just because of AI”

  1. An essay (or any other writing) is about formalizing your thought around a topic. It’s about crafting an argument that can be explored from different angles, and connecting it all to literature in a web of trust. Generative AI can kinda-sorta generate text that almost looks a bit like this if you squint just right, but an essay is about so much more than just pumping out words to hit a required length. I mean, so much essay writing is exactly that. But that’s why we want students to learn to actually do this stuff rather than just deferring to whatever bullshit Clippy squirts into a document for them.

    I’m providing feedback to a bunch of brilliant and intensely creative students, many of whom are clearly struggling with being able to write an essay at that level. It’s not about the content as much as it is about backing up your claims and doing so in a way that the reader can follow and make the leap with you to the creative insight that goes beyond that.

    Finding sources is fundamentally different – for the better – than when we were starting out. Kids these days don’t have to line up to search text in microfiche or microfilm, or to access the CD-ROMs for the ERIC index, etc. They can just plop whatever question they want into a box on Google Scholar, and get high quality results in seconds. But that’s the easy part (and it’s much easier now – GOOD!). It’s the harder stuff about crafting questions, systematically and methodically connecting questions to the literature, making sense of the literature (what’s good? how can I trust it?) and developing a plan to start to answer the questions. That’s what a (good) essay is, and that can’t (shouldn’t) be replaced by tech unless we completely externalize our thinking.

    Says the guy that didn’t cite anything in this mini-essay-of-a-blog-comment.

  2. The sub-title of this essay should be either:
    1.)This Machine Kills Fascists OR,….
    2.)How I Stopped Worrying and Love Control-F
    This is GOOD, all the way down. Search, is the next battle front. Mike Caulfield is onto something with SIFT definitely. Why this essay from back on June 22, says a lot: https://hapgood.us/2023/06/22/the-open-argument-must-be-fed-the-peculiar-case-of-fox-news-and-the-pleasant-smoke/. Emphasis on “Pleasant Smoke”.

    1. Trouble with essays is that they have little practical application in the real world. As an employer I would deliberately pass over graduates for those that could apply practical solutions to problems. And it worked!

  3. “The essay, I’m arguing here, no longer forces students to learn how to research,” writes Dave Cormier. Consequently, “I’m increasingly starting to think that we need to re-evaluate what the basic epistemic skills are that we think people need to make meaning with all this abundance and all the GenAI out there.” It’s even worse than that. Cormier describes a process where he would “write down a bunch of possible articles/books on a piece of paper, go around the library and find said resources, settle in at a table to go through them. I had to read them.” That’s what I did. That’s what I still do, only digitally (that’s where OLDaily originated). But today people don’t read the essays, they do a ‘literature search’ and pull out a set of articles essentially at random based on a keyword (or AI-supported) search. That becomes ‘research’. Now, I don’t think we ‘make meaning’ as suggested by Cormier. But I do suggest that the process is rather more than just search. But if educators don’t do that, they can’t describe it.

  4. I stumbled across ‘It’s all about the search’ about twenty years ago. I was teaching first year performance design students a module called ‘Visual Vocabulary’ which introduced them to ten main art ‘isms’ ( realism, Impressionism, surrealism etc.) The assessment was officially supposed to be a compare and contrast different isms essay. Instead I set them a ‘find and explain your choice’ . Find relevant examples of art, architecture, film, theatre and fashion for each of the isms and explain your choice. They could choose any form of presentation.

    That worked a treat. Students had obviously sometime in the library and online finding examples, choosing what they thought the best, then explaining their choice usually in writing but sometimes graphically. Then choosing the best way to present it all. Most chose a folder-type presentation, but we also had some truly ingenious and creative presentations.

    I suspect Ai , with the right prompts, could do most of that now except the actual presentation.

    1. Mistake mistake mistake. Any argument that says that the AI covers it ignores decades of learning science research.

      1. Would you care to expand on your comment? I’m not a huge believer in much of the learning science that I’ve seen, but am always interested in a new point of view. Your comment, however, doesn’t provide any reason/links for your disagreement.

  5. Dave, generally I agree with you. But search is something more than a good search string. What you did in the library, reading different materials, and synthesizing them, was what was important. Essays required you to do that, but as you point out, that no longer works. Yet, I still think we need Harold Jarche’s PKM (seek – sense – share). What’s important to me is the middle step, making sense of things, after searching and before sharing. If not writing, then what? Not that I have an answer. I’m wondering about diagramming, tho’ will AI be able to do that? Right now my colleague Markus Bernhardt argues that AI can’t do text *and* appropriate images (one or the other). Maybe that’s the trick? Do we need to go back to individual interviews to accompany assignments? I don’t know the answer, but while having good search is important, it’s what you do with the results that also matters.

    1. Yeah, what I mean by ‘search’ is more than simply putting together a string of words, I probably should have expanded more on this here. It’s definitely sensemaking that I’m talking about. meaning making, even (though Stephen disagrees with me in another comment here). Interviews work, though they are tough to scale across what HE has become. My focus, i think, going forward, is not the one special flower who is going to be a genius somewhere, but the rank and file, who are studying the thing I’m talking about not because it’s their absolute passion, but because they’re vaguely interested. That’s all of us, for most things. How do we help the average person, thinking about the thing they are vaguely interested in, make good decisions?

  6. If AI is able to cover the listed eight skills essays teach, then we can probably assume that students won’t be using those skills in the workplace either. What will they be doing instead? Teach those skills.

    If only the people at the very top of an org chart are now going to be engaging with particular kinds of information (because they will tell AI to do what lower level employees previously did), then that information should only be covered at the postgrad level (and perhaps not even then).

    1. I’m suggesting that AI means that those skills aren’t developed using essays, not that it is able to cover the skills. I still think we need those skills.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Creative Commons License
Except where otherwise noted, the content on this site is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.