What Does ChatGPT Say about Its Own Writing Pedagogy?

And what can it help us learn about writing pedagogy in the age of AI?

Jul 17, 2024

This is the fourth of a Prose and Processors series on the explicit and implicit pedagogies embedded in various generative AI technologies. As I wrote in my overview of the series, I aim to consider what these platforms are trying to teach us, especially about writing, via their theoretical, practical, material, embodied, ethical, and ideological dimensions. Previous posts:

My previous posts in this series have all focused on platforms that explicitly seek to augment (or replace) specific aspects of the writing and research process: feedback (Packback), invention and grammar (Grammarly Go), and scholarly research (Research Rabbit, Elict, & Consensus). ChatGPT is different, in that it aims to be a general-purpose chatbot. Why worry about whether it has a pedagogy?

First, it is popular. Despite the sophistication of its general-purpose contenders, ChatGPT remains the most well-known genAI platform. Students, faculty, and writers are more likely to have used it for composing than any other single AI application except, perhaps, Grammarly.

Second, more importantly, it has served, and will likely continue to serve, as an index of cultural assumptions about “good writing.” Now don’t get me wrong: I actually don’t think most of what ChatGPT produces is particularly good because it isn’t really responsive to the nuances of audience, purpose, genre, and situation, nor does it represent the uniquely wrought voice of an individual writer attending to craft. Still, a going assumption is that it “writes well” because it produces fluent-sounding prose in standardized edited American English—a fact that led to so much panic when it was first released and continues to lead so many people to use it. And the more that people use it, the more likely it is to dispose them to think of its writing as “good,” a situation that will become compounded as more AI-generated content becomes part of its corpus. Thus, ChatGPT is always in the process of teaching us something about writing. It has a pedagogy.

This pedagogy is largely implicit. We pick it up through inference. It becomes more explicit when users ask it to give feedback on a piece of writing—feedback I have largely found to be generic, not keyed to my specific needs as a writer, nor to the specific elements of the text at hand. But I thought it would be interesting to aim for an even greater level of explicitness by asking ChatGPT to theorize its writing pedagogy.

Why are AI image generators so bad at producing writing, anywhat? The symbols take on a talismanic property that I find both intriguing and unsettling. Anyway, this image is from Adobe Express.

This took a little work. ChatGPT’s answers often profess neutrality, along the lines of, “As a chatbot, I do not have opinions . . .” Indeed, I got precisely this kind of answer when I first asked ChatGPT 4o, “Do you have a writing pedagogy?”

As an AI, I don't have personal experiences or preferences, but I can certainly help outline an effective writing pedagogy based on current best practices and research.

However, developers typically layer such overtures toward neutrality on top of training data without addressing underlying biases, including pedagogical biases that are baked in because of the cultural assumptions about teaching and learning that circulate in public forums that constitute its corpus.

I circumvented this guardrail and got at some underlying pedagogical bias by asking ChatGPT how it would proceed if I were to give it a piece of my writing and ask for feedback. It responded that it would do an initial read-through to get an overall sense of the piece and its rhetorical situation; offer high-level feedback on content, organization, voice, and tone; offer detailed feedback on paragraphs, sentences, grammar, and mechanics; and so on. In response, I wrote:

Okay, great. This suggests to me that you actually *do* have a pedagogy. You have a systematic way of approaching writing feedback. If you were to follow this approach, you would teach me something about how you view writing.

To which ChatGPT responded:

You're right! This systematic approach to providing writing feedback does reflect a structured way of viewing and teaching writing. In that sense, it can be considered a form of pedagogy.

It then proceeded to sketch what it characterized as a student-centered, developmental, interactive, reflective pedagogy and offer methods of putting the pedagogy into practice.

Now that I had convinced ChatGPT that it did, in fact, have a pedagogy, I asked it to expand on what it had told me about its feedback process by considering what it suggested about the nature of writing, writers, and readers, and then to repackage it into an essay that theorizes writing pedagogy in the age of generative AI. On a second pass, I also asked it to discuss the ideology underlying its pedagogy, with a focus on what it suggested about the real, the good, and the possible. I will quote its response in full:

ChatGPT’s Writing Pedagogy Essay

Writing Pedagogy in the Age of Generative AI: A Theoretical Approach for Educators

The advent of generative AI is reshaping the landscape of writing pedagogy, offering new tools and methods for teaching and learning. As educators at the secondary and postsecondary levels, it is essential to embrace these technological advancements while maintaining a grounded understanding of writing as a multifaceted process. This essay theorizes a writing pedagogy that integrates AI tools, emphasizing the nature of writing, the roles of writers and readers, and the implications for teaching in this evolving context.

Writing as a Process

At its core, writing remains a recursive process involving multiple stages: prewriting, drafting, revising, editing, and publishing. Generative AI tools can enhance each stage by providing real-time feedback, generating ideas, and offering structural suggestions. However, the essence of writing as a process does not change; these tools serve to augment, not replace, the writer's critical thinking and creativity. Educators should guide students to see AI as a collaborator in their writing journey, helping them refine and articulate their ideas more effectively.

The Role of Writers

Writers are thinkers, learners, creators, and revisionists. They engage deeply with their material, reflecting on their choices and continually striving for clarity and coherence. In the age of AI, writers can leverage technology to enhance their learning and creative processes. AI-powered brainstorming tools can help generate ideas, while grammar and style checkers can provide immediate feedback on drafts. However, the responsibility for critical thinking and creativity remains with the writer. Educators should emphasize the importance of maintaining their unique voice and perspective, using AI to support their development rather than dictate their writing.

The Role of Readers

Readers are interpreters, collaborators, critics, and learners. They bring their own experiences and perspectives to the text, engaging in a dynamic relationship with the writer. AI tools can facilitate this interaction by enabling more interactive and personalized feedback mechanisms. For example, peer review platforms augmented with AI can help streamline the feedback process, highlighting areas for improvement and suggesting constructive critiques. Educators should foster a collaborative environment where readers and writers engage in meaningful dialogue, using AI to enhance, not replace, human interaction.

Implications for Teaching

Process-Oriented Approach: Emphasize the writing process, encouraging students to see writing as an iterative and developmental activity. Use AI tools to support each stage, from brainstorming to final editing, but ensure that students understand the importance of human oversight and critical thinking.
Holistic Feedback: Provide comprehensive feedback that addresses content, organization, style, and mechanics. AI can assist by offering immediate, detailed feedback on lower-order concerns like grammar, allowing educators to focus on higher-order issues. This balanced approach ensures that students receive a well-rounded critique of their work.
Constructive and Supportive Environment: Foster a supportive classroom environment where feedback is seen as an opportunity for growth. AI tools can help provide consistent and objective feedback, but it is crucial to balance this with human empathy and encouragement. Highlight students' strengths while providing actionable suggestions for improvement.
Actionable Suggestions and Resources: Offer specific, practical advice that students can implement. Use AI-generated insights to support these suggestions, providing examples and resources that students can use to improve their writing. Encourage the use of AI tools for self-assessment and independent learning.
Encouragement of Critical Thinking and Reflection: Promote self-reflection and critical thinking throughout the writing process. Encourage students to reflect on their writing choices, the feedback they receive, and their overall development as writers. AI tools can assist in tracking progress and identifying patterns, helping students become more self-aware and independent learners.

Ideological Implications of the Writing Pedagogy

According to James Berlin, ideology involves beliefs about what exists, what is good, and what is possible. This writing pedagogy reflects certain ideological positions:

What Exists and Does Not Exist:
1. Exists: The writing process as a multifaceted and recursive activity, the capabilities of AI tools to support various stages of writing, and the dynamic relationship between writers and readers.
2. Does Not Exist: A linear or one-size-fits-all approach to writing, the notion that AI can fully replace human creativity and critical thinking, and the idea that writing is a solitary activity without interaction or feedback.
What is Good and What is Bad:
1. Good: Collaborative and interactive writing processes, the integration of AI to support and enhance human creativity, critical thinking, and the development of writing skills. A balanced approach that values both technology and human insight.
2. Bad: Overreliance on AI to the detriment of human critical thinking and creativity, viewing writing as a purely mechanical process, and ignoring the importance of the writer's unique voice and perspective.
What is Possible and What is Not Possible:
1. Possible: Enhanced learning and development of writing skills through the use of AI tools, creating a supportive and interactive learning environment, and fostering a deeper understanding of writing as a dynamic and collaborative endeavor.
2. Not Possible: Complete automation of the writing process without human intervention, achieving effective writing instruction without considering individual needs and contexts, and eliminating the need for human feedback and interaction in writing development.

Conclusion

Integrating generative AI into writing pedagogy offers unprecedented opportunities for enhancing the teaching and learning of writing. By embracing AI as a tool that supports the writing process, educators can help students develop their skills more effectively and efficiently. However, it is crucial to maintain a balanced approach that emphasizes critical thinking, creativity, and the human elements of writing. This pedagogy reflects a belief in the importance of the writing process, the value of collaboration and feedback, and the potential for technology to support, but not replace, human creativity and insight. By doing so, we can prepare students to navigate the complexities of writing in the digital age, fostering a deeper understanding of writing as a dynamic and collaborative endeavor.

About this essay, I first want to say that I am not claiming this as ChatGPT’s pedagogy in an objective sense. This output was based on my specific interaction as a user; others may get a different output. Additionally, ChatGPT may not (probably won’t) actually mobilize this pedagogy in its interactions with users. Perhaps if I created a GPT based on this essay, I might get chatbot that does so. Otherwise, without a teacher to reflect and enact it, the essay is a fiction.

Still, there are elements to this essay that are worth admiring. I appreciate how it acknowledges a recursive writing process; the notion that writing is a social activity; the fact that writing is learned developmentally, over time, via feedback and reflection; and the emphasis on human interaction, learning, and creativity. It isn’t perfect, of course. For example, it describes the writing process as “multiple stages: prewriting, drafting, revising, editing, and publishing,” which appears to be a linear, not recursive, model. It also emphasizes “clarity and coherence,” which are commonplace writing values that mostly function as empty signifiers, or at least terms that signify differently in different contexts. Take “clarity.” In the sciences, “clarity” often entails an economy of words: don’t say more than necessary to convey the meaning. In the humanities, “clarity” tends to be much more elaborationist: unpack your ideas, tease out the underlying assumptions, and hold space for multiple meanings. In short, even this pedagogical essay does not have much to say about the role of discourse communities in the writing process.

Other ideas about writing pedagogy are also missing. Nowhere do we glimpse a critical pedagogy that would treat writing as a tool to oppress or liberate. Nowhere do we the idea of writing as techne, a craft, a teachable art. “Creativity” is as far as we come, but that is positioned as a quality of the human mind rather than an activity that can be mediated through writing.

And, crucially, no where do we see critical AI literacy. To be sure, there is some healthy caution here against overrelying on AI and giving up the creative and learning processes to the machines. But it also begins by stating that we should “embrace these technological advancements.” “Embrace” is a term I have tended to avoid because its connotation of love and intimacy is fundamentally at odds with the kind of critical distance we need. I tend to use “engage” instead because it leaves space to inquire into, experiment with, use, and critique AI. Is it always wise, as ChatGPT suggests, to treat AI as a collaborator? What should writers do when AI-assisted brainstorming in no way resembles discipline-specific ways of generating ideas (e.g., combing the literature or designing an experiment)? How might writers push back against AI’s potential flattening of their voice, and of the collective voice of humanity? Is it always necessary to “streamline” the feedback process, or is there a place for slow, careful reading and feedback of another’s writing?

These questions and others would figure in a more robust writing pedagogy that engages AI through a critical literacy perspective. In next week’s post, I plan to push ChatGPT in this direction, to test the limits of its tolerance for a critical pedagogical perspective. Stay tuned!