I’ve limited myself to one hour of writing for this post, so it’s more a collection of thoughts than a coherent narrative. I expect I’ll have lots more to say on this topic in the future. For now, I just want to get the beginnings of this idea out into the world, together with some initial implications.
Since ChatGPT’s meteoric rise to popularity, I’ve constantly been amazed by the creative ways people have imagined to make use of generative AI. For months now it feels like each week I read about some new way generative AI is being used that just completely blows my mind. For example, my son who is studying cybersecurity recently told me prompted ChatGPT along these lines: ‘You are the hiring manager for a cybersecurity position at a large company. Conduct a mock interview with me, asking me ten conceptual questions. After I have answered all the questions, give me feedback on my answers. Then ask me ten technical questions. After I have answered all the questions, give me feedback on my answers again.’ My son tells me this exercise was incredibly helpful and he plans to do it several more times before having a “real” interview.
By definition, our lack of imagination is the only limit on our ability to use these tools in novel ways.
I continue to try to imagine ways generative AI can impact teaching and learning, including learning materials like textbooks. Earlier this week I started wondering – what if, in the future, educators didn’t write textbooks at all? What if, instead, we only wrote structured collections of highly crafted prompts? Instead of reading a static textbook in a linear fashion, the learner would use the prompts to interact with a large language model. These prompts could help learners ask for things like:
- overviews and in-depth explanations of specific topics in a specific sequence,
- examples that the learner finds personally relevant and interesting,
- interactive practice – including open-ended exercises – with immediate, corrective feedback,
- the structure of the relationships between ideas and concepts,
- etc.
As a very crude first approximation, imagine taking an existing textbook, finding each heading in the text, turning the heading into a prompt, and then deleting all the text beneath the heading. Then do the same with every colorful little box containing an example, case study, etc.
Rather than “reading,” this learning experience would be more like a conversation. Have you ever been reading a textbook, not understood something you read, and wished you could just turn to someone and ask them to explain it in a slightly different way? With a generative textbook you can. You simply ask, “Can you explain it to me a different way?” or “Do you mean (restate your understanding)?” or “Could you use a different example – maybe one using (your hobby or interest here)?”
Imagine how much more natural it would be to teach metacognitive skills, information literacy, and related topics when a learner’s primary activity is asking questions of an LLM, rather than reading a static text. Learning to ask useful questions – whether of an LLM, another person, or the universe itself – is directly at the center of the educational enterprise.
In graduate school I was endlessly fascinated with Gordon Pask’s conversation theory, which seems to have broad implications for the design of this new kind of instructional material. I think the world is about to rediscover the work of Pask and Pangaro and others.
A structured collection of specific prompts bears very little resemblance to the textbooks we all used growing up. However, one of the best ways to help drive the initial adoption of new inventions is to couch them in the language of the old – think “the horseless carriage.” I’m sure we will come up with a more appropriately descriptive name at some point in the future for this kind of learning material. But for now, by calling them “textbooks” I think we will greatly increase the likelihood of instructors actually considering adopting them. I would call these learning materials something like “conversational textbooks” (with reference to Pask’s work) except that phrase already means something different. So for the time being I’m going to call these “generative textbooks.”
As an initial first step in this direction, generative textbooks could be created as supplemental companions to existing static textbooks. However, I predict many students will quickly abandon the traditional format textbook in favor of the more interactive, open-ended, and personalized nature of the generative textbook.
And for those of you who expect every post on this blog to be related to OER in some way (you did notice I changed the name of the blog, right?), yes, you could openly license your generative textbook (collection of prompts). But the fundamental role of OER changes dramatically in future scenarios where the majority of the learning material a student engages with is generated on the fly by an LLM, and was never eligible for copyright protection in the first place.
In the case of generative textbooks, prompt engineering is learning engineering, or instructional design, or whatever they call it in your world.