Over the last decade or so, the concept of “metacognition,” or thinking about one’s thinking, has moved from academic conversations in the fields of psychology, cognitive science, and education into mainstream conversations about learning.¹ It’s clear: teaching students to be aware of and reflect on their cognition at different points in their learning process is connected to higher academic performance and long-term retention of skills and knowledge.²

Since early 2023, discussions about Artificial Intelligence (AI) have dominated educational conversations, with great attention given to ChatGPT. Here at CDIL, we’ve been talking with faculty members across BC schools and departments about their concerns, questions, and hopes for the way that this quickly-changing technology might or is already impacting their classrooms. 

In my own pedagogical work as a writing and literature instructor, I’ve experimented with AI in my classroom. What I find most exciting about this technology is the way it surfaces existential questions about what it means to think, what it means to create knowledge, and what it means to be human. 

One approach we as instructors can take is to invite our students to engage with the non-human, alien brain of ChatGPT in a way that deepens their understanding of their own cognition.³ Explicitly prompting students to think about their thinking in relation to non-human cognition prompts them to not only develop a sense of metacognitive awareness, but pushes them to do this work from a distinctly humanist lens – juxtaposing their cognition with that of AI helps highlight their own humanity while centering the value their cognition brings to their work. 

Steps for Metacognitive Inquiry

One way that you can invite students into this metacognitive inquiry is as follows:

1. Have students use ChatGPT (or another form of generative AI) as part of an assignment to generate writing. 
2. Then, direct students to annotate ChatGPT’s output, analyzing the writing with particular attention to its strengths and limitations.
3. Next, ask students to revise the ChatGPT generated writing. 
4. Finally, have students annotate their own work, explaining their revisions and justifying why those revisions are stronger. 

Following these steps creates a subtle yet powerful shift in the way that writing is used to demonstrate student learning. The pedagogical move of asking students to annotate both the ChatGPT generated and their own writing shifts the focus from the finished written product to students’ annotations – the written evidence of their metacognition and their analysis of the “alien” cognition of ChatGPT. 

Takeaways from My Classroom

I followed the above process in my Literature Core class, “Bad Girls: Unruly, Cruel, Nasty Women in Literature, Film, and Popular Culture,” in Spring 2023 with prompts that asked students to perform a close reading on assigned texts for the course. 

After students completed this assignment, I administered a survey asking for their feedback on the activity, including their thoughts about how, if at all, it shifted their thinking about what it means to write. 

Several themes emerged across their qualitative responses, but by far, students focused on how this assignment sparked them to think about their own thinking. One student, in particular, referred to ChatGPT as a “sort of conversation partner,” continuing by discussing how thinking about her own thinking in relation to AI helped her to clarify her purpose and make her argument stronger. 

Assignments like this, where the students’ metacognitive work on their own writing, rather than the writing product itself, not only anticipate and eliminate any issues with unraveling the difference between AI-generated versus human-produced writing but they invite students to question what it means to think itself. By creating an environment that necessitates engagement with others– both human and nonhuman– in the production of knowledge, we’re equipping our students with a deeper understanding of their own cognition and of their own humanity. 

Have you used AI as part of an assignment for your course? Let us know by contacting us.

Footnotes

1. “Metacognitive knowledge” was first defined by John H. Flavell in “Metacognition and cognitive monitoring: A new area of cognitive-developmental inquiry” (1979) as “one’s knowledge concerning one’s own cognitive processes and products or anything related to them, e.g. the learning-relevant properties of information and data” (233). Since the beginning of the 21st century, studies on the connection between academic performance and metacognition proliferated, with the first issue of Metacognition and Learning published in 2006.

2. For more on the connection between metacognition and academic performance, particularly its role in developing critical thinking, see Hanley, G.L. 1995; Kuyper, H., M.P.C. van der Werf, and M.J. Lubbers 2000; Lehmann, T., I. Hähnlein, and D. Ifenthaler 2014; Venman, M.V., B.H. Van Hout-Wolters, and P. Afferbach 2006; Vrugt, A., and F.J. Oort 2008. 

3. My thinking in this area is theoretically grounded in the rich inquiry currently stemming from the new materialisms. Specifically, I’m inspired by Karan Barad’s concepts of “agential realism” and “intra-action” in Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning.