LLMs in Education

In December, I attended a Women in Tech retreat hosted by Sharon Weaver of Smarter Consulting . One of the participants at the retreat has spent her entire career in Education, and she wanted to talk to me about AI and the present and future impacts it will have on Education. For the purpose of this article, when I refer to “AI,” I am only referring to Large Language Models (LLMs).

The first question I asked was, “If all students are home schooled, what are they missing out on?” The response was immediate and what we’ve heard from educators for years – socialization. Okay. So, think about that. What are we really trying to do when we made kids read a book? We are socializing them with the curriculum of previous generations. And by doing so, we lay a literary foundation for the next generation.

Were we trying to expand their minds? Sure, for some, but what we accomplished over the years was getting kids to do what they were told (a crucial part of socialization) and to make sure that most kids had all read the same literature. And curriculum is quite a hot topic. Who decides what books students read? Who decides how they are socialized? I believe a lot of that conversation can be aided by AI. If you want a more diverse reading list, it’s only an LLM prompt away. But that’s not the real issue, is it? The issue is getting that curriculum approved. And more importantly, standardized so we can test on it.

So, yes, we are socializing them but it’s not just about teaching them to read and write, is it? Isn’t is also about preparing them to be employable citizens? Hold that thought.

What would happen if our kids never read a book again? What if they only read summaries? What’s the consequence of that? I mean really. What’s the consequence? As adults, we have podcasts and apps that do just that – summarize books for us or read them to us. Just like I will always have a calculator and the internet, I will always be able to find a summary of a book online. That was happening way before AI. Anyone remember Cliff’s Notes? So, other than ensuring that a child knows HOW to read (because that’s still important), what else were we hoping to accomplish? What was the point of the book report?

In addition to making sure that our children learn how to write, I suspect it was also teaching to the standardized test of reading comprehension. That’s a skill they need to get a good score on a standardized test that will get them into college. And this is where I think it starts to get scary for educators and their administrators. Are we moving towards a world with no standardized tests that doesn’t require college degrees?

I propose that AI will see a rise in the number of students who are able to successfully obtain college degrees. Imagine the time savings if I didn’t have to read that textbook cover to cover? What if I could ask AI to give me sample test questions based on the syllabus and textbook? It would mean that more students could cope with the stress of consuming a lot of content across multiple subjects in a shorter amount of time. What if I could graduate faster?

Here's what we do know. This technology is not going away. Just like we tried to keep cell phones out of the classroom, AI will have to be accepted as part of the learning process and incorporated into it. I think the challenge for educators and their administrators is figuring out how to embrace that change sooner rather than later. I think that colleges with a clear strategy are going to be competitive against the latent adopters in their industry. I think embracing AI should become a part of their recruitment strategy, “We use AI to prepare you for the future.”

So the real question is, if the real value of in-person education is socialization, then how do we accomplish that using AI? Education (in fact, all industries) will need to prioritize what user experience they think is key to accomplish their long-term goals. If I was going to use AI to socialize someone, I’d focus on these three things:

• Creating equity in the classroom by using AI to accommodate different learning styles. There are always going to be students who love reading and writing. AI will help them differently than a student with dyslexia.
• Broadening curriculum to include content that looks more like the individual student. AI can reduce the burden on educators to broaden curriculum. The bottleneck will be a district’s ability to quickly approve the new content. Here is where automation can play a role – what if students had a way to nominate what book they want to read, and we automatically approve anything on a list that keeps growing with each nomination to stay relevant?
• Help students prepare for their future by giving them more access to opportunities. AI may create more opportunities for students to be successful in college. It can also help those who would rather be artists. Or both!

If the “real” goal of education is creating employable citizens, LLM’s are far-better equipped to do that than regionally focused information. If a kid doesn’t want to grow up to work at the plant down the road, they need to know what else is possible. Education has been known to stifle potential simply by being ignorant of the rest of the world. People with access to LLMs will do more and they will accomplish it faster. We need to get Education into a mindset of eliminating mundane tasks to free students up to do more innovative work.

?