The other day my friend David Cutler posted a short video that demonstrated with crystal clarity the possibilities of AI as a tool for composition, and a colleague shared with me some of their own experiments. It works really well, possibly TOO well if you worry that it will just become a tool for academic dishonesty and general short-cutting of the learning process. But great minds at great universities seem already to have developed apps for spotting AI-generated text, so perhaps that menace may be brought under control, sort of, at some point.
But David’s excellent video inspired my inner optimist to follow a line of thought that I had been musing on for a while already: How to make a virtue of necessity; how to find ways to harness the benefits of AI’s power to stimulate, rather than replace, student thinking and learning. In the video, David challenges an AI engine to generate some essays on a text, his prompts sufficiently nuanced to require that the AI demonstrate its capacity (Or is it an ability? How far may I anthropomorphize here?) to draw upon important textual content and even specific elements (page numbers, no less!) of the book under analysis as it creates, almost instantly, very well-crafted essays in response to the prompts. Yikes!! The thing is good—really, really good.
It occurred to me, and I am shamelessly going to quote and paraphrase here the comment I left on David Cutler’s video post, that perhaps a fruitful and and affirming approach to this awesome new technology might be to focus more on the way we approach the matter through the development by human students of questions and prompts—of the disposition to and skills of inquiry—and less on AI's ability (there; I did it) to compile astonishingly competent answers from the gazillion human-created sources upon which it can draw. What if a teacher were to assign students the task of crafting, oh, say, three very different essay prompts for the AI engine to answer, with the point being that the responses would illuminate the major themes of the work and the ways in which the author explores and expresses these themes? That’s kind of the easy part. The AI bangs out its three lovely essays, voilá! The student has posed questions and can now read expert responses that will likely contain connections and nuances the student may not have previously noted. A learning experience in itself, this!
BUT THEN the task of the student, hopefully in a medium that is somehow “AI-proof” (some have suggested that making students handwrite work as a way to circumvent AI, but that’s a whole other discussion), would be to reflect on the response essays in relation to the prompts, on the value of the prompts themselves in light of the responses, and then on their own learning in the process. This work, accomplished in whatever ways seem apt—in class? oral? workshopped collaboratively in small groups?—would in itself provide quite a bit of illumination on any text, historical era or event, or other cultural artifact/production under study.
In creating such clever artificial intelligence engines, we—and once these things are ubiquitous it’s “we,” not “they”—have created and are unleashing a tool with enormous new and as yet not fully understood power. We’re educators, and we need to find ways to apply this kind of power and not simply spend our time trying to thwart its misuse—though this must also be on our minds. By asking students to reflect on what “AI says” relative to their own understandings, we can create critical dialogues instead of resigning ourselves to the imminence and inevitability of a “singularity” in which machines subsume or become us.
Our human task, as educators who have committed ourselves to helping create a better future by helping to expand the perspectives and hone the habits of mind of rising generations, is to think deeply and clearly about what we ask such powerful tools to do and how we ask them to accomplish the tasks we set for them—not just by assigning constraints and parameters but even, very intentionally, engaging with the real stuff of life: emotions, ethics, and larger concerns like equity, justice, and human survival.
AI is great power, and we have to learn for ourselves and to teach our children how to use such power for positive, humane, moral ends.
For more on this topic, check out this accompanying blog written by our Senior Director, Sarah Hanawald: AI Can Write Essays: What Does This Mean for Educators?
Don't miss our weekly blog posts by joining our newsletter mailing list below:
Brad Rathgeber (he/him/his)