I’ve had an AI policy on my syllabus for three years and it has mostly remained the same. I assign a lot of personal scholarly essays, so for the most part, I wasn’t seeing much AI use in student essays (at least not in a problematic way) until this past Spring 2026 semester, when many of the cover letter assignments started sounding the same. That assignment still works–more on that in another post but first I just need to cry about it for a second. I got a batch of vague cover letters detailing the same boring goals and lacking in personality. The use of AI to generate students’ own personal goals was clear as day. Felt like a knife to the heart that students were not centering themselves in their learning even when I designed an assignment for that purpose. 😭 💔 My next post will be about how I’ve changed it.

AI has gotten better, it’s being used more often, and students are generally being told whether it’s allowed or not without being given the mentoring they need and deserve. My very short and basic policy was not helpful. I spent all summer ruminating and developing ideas that I will share over the course of three blog posts. This post focuses on my revised AI Policy.

A couple things to disclose first: (1) the course meets in person, has about 25 students, and is a writing intensive with mostly non-majors; (2) my policy is informed by James Lang’s beautiful book Distracted (published before GenAI hit higher ed yet more relevant than ever) as well as being in community with colleagues (thank you Jack Kenigsberg, Raj Korpan, and Shiao-Chuan Kung for Camp AI); and (3) I created a chatbot to support my students’ revision process especially for those writers who need more support than I have capacity to provide–that might irk some people but stay with me to the end–don’t skewer me just yet!

My revised AI policy reflects a classroom where GenAI is sometimes OK and sometimes not OK. Lang’s take on device policies is that there is little benefit to a wholesale ban or adoption of technology, and he encourages instructors to set the boundaries instead of having students do this part from scratch. We as instructors have to set the appropriate and healthful conditions for deep and meaningful learning to happen. My goals in writing this policy were to make clear when to use GenAI and when not to; focus on learning most of all; and to write in a voice that is warm and thoughtful to build connection and trust with students. The policy will be reviewed by students. I’ll invite comments on the shared doc, answer questions, and revise it based on student feedback.

Feel free to use all of it or parts of it, adapt it, recycle it...Note: at the bottom of the policy are “Steps for Essay Revision” I’ve outlined for students so they have a clear blueprint for what to do when.

My Revised AI Policy (as it appears on my Fall 2026 syllabus)

When you use AI in your process matters. The point of being here is to learn, discover and strengthen your authentic voice in writing, and to develop your skills for a self-directed future that is meaningful to you. When you use GenAI first, to do the initial work for you, you are not engaging in the organic, wild, messiness of growth that only you can do. You cannot use GenAI to do your work for you in this class. If you put an assignment into GenAI and ask it to produce a response first, even if you tweak it later, that is not the same as producing something you own (in your brain) forever, something you developed from scratch in your own voice. Using AI to generate even a first draft is self-sabotage and off-limits in this course. Hard no. Don’t do it. 

Similarly, using GenAI to get your first answer to a question deprives you of social, collaborative learning—the most deep and meaningful learning we can engage in. Please consult the syllabus first; if your question is not answered, ask a classmate; and if your question is still not answered—or you are uncertain—then email your instructor. If it’s a general question about the world (not specific to our class), then you would do better to consult Google and/or Wikipedia and verify the accuracy of what you find. Tools like chatGPT are Large Language Models (LLMs) that will not give you an answer the way a calculator would. A calculator operates on the same mathematical logic that humans do. LLMs do not. LLMs use probability models that give you the most likely response meant to please the user and not to be the most accurate. Have you ever encountered a truth that was difficult or surprising? Learning is sometimes painful and difficult. GenAI was built to relieve that pain. 

However, there are productive ways to use AI, later in your process, that are allowed in this course. If you don’t understand the context for a text, you can ask Google (Google is AI) to find you a summary, and click through the links it gives you, in addition to reading the text yourself. This could enhance your reading experience and depth of understanding. 

If you’ve already written something and want to know how you could develop or polish it more, or if you’re concerned about your grammar or paragraphs (e.g., topic sentences, transitions), try asking different AI agents for feedback. In fact, I built an agent for this specific course and I prefer you use that one, in our closed university Microsoft system, because it is substantially safer and tailored to our course. The quality of your ideas and demonstrating your understanding of the texts we read should remain the same. AI can help you with revisions to identify areas for improvement in the writing. You can use it the same way you might give a written draft to a peer and ask them for help. Then analyze your original work. Does the feedback make sense? If so, what can you learn from the difference it makes? Always look at the end result and ask yourself, “Does this sound like me?” (Clarification: your “voice” in writing does not mean using conversational tone and run-on sentences like a talk-to-text situation; you still need to edit and revise your ideas into complete sentences and thoughts organized into paragraphs that communicate your ideas clearly.)

When you use AI, you must acknowledge AI use, disclose this to me, up front with a brief process statement and a summary of your interactions with the course-specific chat bot. A process statement summarizes where and when in your process you used AI, why you used it, and what worked well and what didn’t work well. Please include a citation in the essay itself if you incorporate AI-generated text, and submit the document with a brief process summary (3-4 sentences) in the comments on Brightspace. You must download a chat summary and submit that to me as well. 

When using the resources available to you, do so to learn. You owe that to yourself. This is your education and, ideally, you leave this course owning knowledge and skills that belong only to you. In addition, be critical of your sources. Do not be fooled: chatGPT is not an author (see Modern Language Association (MLA) AI Policy). ChatGPT owns whatever you put into it. Before you use chatGPT, consider how it has been built: “OpenAI Used Kenyan Workers on Less than $2 Per Hour to Make ChatGPT Less Toxic.” I strongly recommend using CoPilot via your university Outlook account instead because at least your data is better protected. 

Steps for Essay Revision

You may revise ANY essay for this class for a better grade—ONCE. If you are given a “Revise and Resubmit” in my feedback on your essay, which you will be able to see in Brightspace, that means you not only have the option but also my encouragement and sincere hope that you will read my feedback and revise and resubmit the essay for a better grade. Revisions must be substantial—not just surface-level. Here are the steps you should take to revise your essay: 

• First, study P.I.E. paragraphs, work on your topic sentences, integrating quotes, explaining them, and focus on organizing your ideas effectively so you have room to go in depth and detail. 
• Next, make an appointment with the Writing Center for help developing your essay.  [Link omitted] If they are fully booked, ask them for resources! Then try swapping drafts with a friend for peer feedback.
• Finally, once you have a solid working draft, you might try this course-specific chat bot for polishing your essay before submission. [Link omitted]
• Type “END” in the chat and download the chat summary when you are finished with the bot, and submit that alongside your revised essay in Brightspace. Remember to add a process summary of 3-4 sentences disclosing your use of AI.

One Analog Thing To Do

The flipped classroom is more important than ever. Our class time is where we have the most control and where we can (re)direct attention. Let’s carve out more in-class time for the hardest parts of learning.

This semester, I’m giving students small journals, pens, and fidgets, and I’m committing to giving them real chunks of class time for brainstorming essay ideas, and workshopping ideas with a partner. They will generate their first couple of drafty paragraph(s) in class in their journals, which I will collect later when they submit their essays. The journal draft counts toward their essay grade. I want human-generated ideas first. I want them to stare at the blank page first, and struggle to generate something of their own. And I’m going to model it for them, do it with them, struggle together in community.

In the past, I’ve used the free Blue books we can get from the college, but this year I’m inspired by my colleague, Nancy Hightower, who reminded me how much it means to hold something meaningful in your hand, something someone gave you, something that you actually want to open and use. So I invested in a bulk pack of tiny journals. And I’m surrendering some class time for the hard labor of generating first drafts.

It’s a precious gift to be in a room together, to be fully present with one another, and to engage in something truly difficult in community. So, yeah, I created a chatbot and I’ve got a revised AI-integrated cover letter assignment, and I’ll model for them how I use AI occasionally so I can mentor them on good practices, meta-prompting, and verifying results, but I’m also going mostly analog in my classroom because there’s a time and a place…and my priorities are always going to be deep and meaningful learning in community.