The Philosophers' Cocoon

A safe and supportive forum for philosophers.

Owned & Moderated by Marcus Arvan (University of Tampa)

Contact: marvan@ut.edu

Blog mission & moderation policy

IMPORTANT NOTE on how to ensure anonymity in commenting
Job-Market Reporting Thread (2025-26)
Job-Market Discussion Thread (2025-26)

recent posts

about

Teaching writing in the AI/LLM era?

In our new “how can we help you?” thread, a reader asks:

I am curious how other people teach writing in the age of LLMs. I am teaching in a political science department and have to teach students some philosophical writing in order for them to be able to write term papers and theses in political theory/philosophy.

Previously, my strategy would have been to give students a couple of small writing tasks during the semester and provide some feedback on those tasks. However, with what LLMs are currently capable of and with the degree of LLM-use I have seen, I am pessimistic about the prospects of that strategy. My expectation is that anything like “Summarize the argument on p. 2” or “Write a brief response to argument X” will yield a large number of machine-generated responses.

So I would be curious to hear how others are teaching writing now. In particular, I am curious how you can integrate it into a regular seminar for students that might only take a small number of philosophy courses. Do you do writing exercises in class? Do you just stop teaching writing?

I am curious to hear about this too. How (if at all) do you teach writing now?

Posted in

9 responses to “Teaching writing in the AI/LLM era?”

  1. Mike Titelbaum

    I’ve only tried this one semester so far, so I don’t have representative data about whether it worked. But what I did in my most recent Philosophy 101 was: I gave the students two standard, take-home essay-writing assignments over the semester, and an in-class final. The essays were graded, but only worth 10% of their grade each. The final was much more. I told them that the point of the essays was to learn a skill—philosophical writing—that they would need on the final. This would be their only chance to develop that skill and get feedback on how well they had done. If they used AI to write those essays, they would then be writing philosophical essays on their final exam for the first time ever, having received no feedback or chance to improve. Then I left it up to them.

    Reply
    1. Anonymous

      This is smart! Thank you for sharing.

      Reply
  2. Anonymous

    I would be really interested to hear back about how this went, for what it’s worth! I like this as a strategy.

    Reply
  3. Anonymous

    In-class writing seems to be the only sure-fire method of preventing the use of AI. But there are serious drawbacks to that, too, including the fact that the current generation has very little experience writing by hand (and so will do a worse job), in class they don’t have access to sources and revision after reflection (keys to good writing), and the fact that students with learning and other disabilities are disadvantaged in such cases.

    Since we have not been incentivized or given the tools to prevent AI use (quite the opposite in many cases), we shouldn’t spend too much effort trying to prevent it. I explain to students why it is to their disadvantage to use AI, and try to structure assignments in ways that make AI use less likely to succeed. But in the end I still teach writing in the same way I always have. Those who learn from it will be at a big relative advantage compared to those who don’t, which is some consolation for the fact that a significant subset of students are determined to prevent themselves from learning.

    Reply
    1. Anonymous

      Thanks for your response! Could you give an example for how you’d structure assignments for teaching writing in ways that would inhibit AI use?

      Reply
      1. Anonymous

        I’m not the person you’re responding to, but one assignment that has worked well for me is requiring students to formally reconstruct an argument from a reading and object to one of the premises. Possibly future LLMs will improve at this, but for the time being, it seems like LLMs are 1) very bad at giving valid argument reconstructions and 2) usually provide generic, boilerplate, “but this is a vague concept—what is fairness?” style objections. I am hopeful that it at least disincentivizes students to use LLMs when they do poorly on an assignment using them.

  4. Daniel

    I have been using oral exams attached to each writing assignment. Students have to come talk with me about their papers.

    Reply
  5. AGT

    Oral exams and/or supervision during the semester to produce a final essay. Perhaps coupled with a presentation in class also connected it to the final essay and supervision topic. In class exams on computers if necessary, with special allowance for students with disabilities and other legitimate hindrances in writing.

    Reply
  6. Anonymous

    I was quite intrigued by this “Multi day In-class lockdown browser assignment” by John Robison has anyone tried it? https://dailynous.com/2025/04/02/the-multi-day-in-class-lockdown-browser-essay-assignment-guest/

    Reply

Leave a Reply

Discover more from The Philosophers' Cocoon

Subscribe now to keep reading and get access to the full archive.

Continue reading