Photo by Skylar Kang on Pexels

Teaching Writing in an AI World

Pennylyn Dykstra-Pruim and Jennifer Redmann—

Our AI world poses renewed challenges for educators. How do we address AI-supported learning conditions that disincentivize “hard” work in favor of a potentially “easy” route to an acceptable grade? We want to learn from our students and their perspectives, and we want them to build communication skills that allow them to express their perspectives clearly and in their own voices. How do we train learners to use new tools (including those provided by AI) productively while also motivating them to achieve learning goals? How do we persist in this effort in a ChatGPT world? Enter the MPG approach to teaching writing.

Model-based, Process-oriented, and Genre-focused, the MPG approach to teaching writing integrates several best practices into a cohesive pedagogy. The MPG approach offers step-by-step skill development for analyzing and using text models; composing and revising sentences, paragraphs, and full texts; and understanding text genres in cultural context.

The MPG approach relies on models but not just for established genres like a resume or film review; helpful models also guide novice and intermediate-level language learners in building sentences, paragraphs, and full texts. While most educators readily concede the need for a lot of scaffolding (models, tips, clear directions) for beginning language students learning to write, the same applies to those writing in their first language. Recall the first time you put together a resume or a letter of application or a wedding invitation. You looked at what others had done and you found models, some that you put in your like pile and others that you rejected. Likewise, our students can benefit from an integrated focus on models—Which of these three opening paragraphs appeals to you most and why? How does the use of semicolons in example 1 clarify or frustrate the reader compared to the use of other forms of punctuation in example 2? What is the effect of moving this participial phrase from the end to the beginning of this sentence? Which elements in this short story make it engaging? How does your thinking process change if the summary statement is withheld until the end of the paragraph? 

Finding and exploring models is where the internet, Google, and, yes, ChatGPT can prove invaluable. We can find or call forth an opinion piece on global warming; we can ask for a text in the style of a right-wing nationalist, a children’s book, a poem, or Shakespeare. We can ask for help with a good topic sentence that will appeal to teenagers or women or EU citizens. And, most importantly, we can learn to analyze, adapt, or reject those models.

Next, for those learning to write, whether in a second language or in a first language, highlighting a helpful pathway to success—a good process—teaches effective habits and builds skills for a lifetime of writing. For producing a wedding invitation, you would likely follow a process, even if that process relied heavily on “hire a professional.” You would focus on getting the facts right but also consider different fonts and how they work with layout and colors. Regardless of the order in which you choose each element, some steps must be undertaken before others. In a parallel way, “Write a review of the film X” or “Your second draft is due Friday” may be adequate for some students, but most learners benefit from concrete instructions that help guide and direct their energies as they develop their writing abilities. 

A focus on process can also make reliance on ChatGPT to “do all the work” less attractive or convenient. For example, in MPG-based writing pedagogy, we break down a big task (such as writing a film review) into smaller chunks (x, y, and z). Next, we integrate exchange and feedback from teachers and peers. We expect that these edited smaller chunks will eventually be integrated into a final text with endnote comments about why the author chose one version over another. This process orientation asks our learners to produce, reflect, choose, and edit; working with ChatGPT to do each of those steps would likely be more effort than it’s worth. We’re not, after all, teaching a product; we’re teaching a process—a “how to.” This shift of focus from product to process also means that we need to carefully consider our assessment. If we have only been grading the final product and now we aren’t completely sure if that product is written by Chris or by Chris’ (clever?) use of a computer app, then we need to assess the process. Writing is a process and grading the final draft is only one piece of assessing the skill of writing.

Inherent to the creation of every text are assumptions about genre. In crafting that invitation, you would consider what information guests need, but also how best to communicate that information in ways that are clear and beautiful, interesting or whimsical. Once learners have an understanding of text genres, they have some power to play with or play off the cultural expectations associated with that genre. Understanding the components of an effective op-ed or a compelling short story or an engaging song helps us craft our own texts. Asking students to articulate how they lean into or subvert the cultural expectations of genre in their own writing (with side bars or footnotes or a short presentation) again thwarts complete reliance on something produced by a machine. It recenters the process of writing by spotlighting and articulating the individual decisions within that process.

A consistent and robust MPG approach to teaching writing helps instructors design writing tasks with care, from model collection, analysis, and curation through a well-designed, multi-step process of producing, reviewing, and editing a text within the context of textual genres.  If we integrate accountability and multiple steps in the writing process, then the practice that leads to learning becomes the path of least resistance. The final written product retains importance but only as one artifact in a portfolio of assessed tasks. If we can incorporate good uses for new tech tools, teach and assess the writing process, and enable our students to follow and diverge from cultural expectations of genre in calculated ways, our students just might learn even more about how to write and how to talk about their own writing with insight and nuance.


Pennylyn Dykstra-Pruim and Jennifer Redmann are authors of Schreiben Lernen: A Writing Guide for Learners of German (Second Edition), in which the MPG approach to teaching writing remains a central focus. Pennylyn Dykstra-Pruim is professor of German and associate dean for diversity and inclusion at Calvin University. Her previous books include Understanding Us & Them and Christians and Cultural Difference. She lives in Grand Rapids, MI. Jennifer Redmann is professor of German at Franklin & Marshall College. With Dykstra-Pruim, she coedited A Writing Guide for Learners of Chinese. She lives in Lancaster, PA.


Recent Posts

All Blogs

Categories