The Box and the Mirror – Part I
The Future of AI in Education and Education with AI
A generous reading of Umberto Echo’s devastating The Name of the Rose would posit that the purpose of the book was to highlight the first quantum shift in education since oral traditionalists shunned alphabets. The book’s plot, revolving around a detective monk and his young apprentice’s visit to a famous library, highlights the lengths people went to in order to learn, and warns against the dissemination of information.
The birth of the printing press had a number of detractors, it’s true. An Italian Benedictine monk even went so far as to say:
“The pen is a virgin… but the printing press is a whore!”
He even spent a year of his life handwriting hundreds of copies of his Polemic Against Printing in order to prove it.
There have been major steps in globalising education since then, of course. The reformation of the 16th century. The adoption of state mandated education in the 19th. The decentralisation of government approved curriculum. The inclusion of minorities - and in the case of women, majorities - within equal access to education. The advent of the personal computer. The overhead projector. Flipped classrooms. Scribe by the side. Experiential education. Metacognition...
The list of these advancements are endless. To a large degree, however, these are simply stones, bouncing across the water in ever-decreasing arcs from the originator that allowed knowledge to be transmitted to the masses in the first place. Since the adoption of the printing press, nothing has had such a profound effect on education.
Until now.
It feels like just the blink of an eye since the world was introduced to AI via the large language models (LLMs) of ChatGPT and Bard, and we’ve barely had time to start to think about their impact on education. But, the International Artificial Intelligence in Education Society (IAIED), an international interdisciplinary group that publishes regularly in journals, comprised of some of education's leading minds, has been meeting since 1997. And they're not alone. UNESCO produced a thorough report on the emerging role of AI in education in 2022, and in 2020 over one thousand AI apps were already freely available, tutoring our children and guiding them through exam-based curricula.
Some early responders have been quick to jump on the bandwagon, but most educational institutions are still far behind on the bell-curve of change, with many burying their heads at the back, refusing to see AI as anything more than an opportunity to cheat. For those of us old enough to remember the Microsoft paperclip, the reaction of the education community shouldn’t surprise us. And those of us even older, old enough to be around at the dawn of the word processor, remember the furore that accompanied it, coupled with the inevitable creep that led to students and teachers forgetting how to spell because auto-correct is everything. For those of us whose memory stretches beyond that, we still remember being taught to type, rather than coming at it as a natural response to the world around us. Teaching touch typing to a twelve year old would be like teaching a fish to swim these days.
So it’s not surprising that the education society at large, while accepting that change is inevitable, think they have time. 72% of educators believe that AI will positively affect the learning environment. 63% understand that it will replace their workforce to an extent, but the talk in the education world is still largely based around supplementing, rather than replacing the workforce.
That’s probably up there with the people still buying houses in the shadow of a rumbling Vesuvius.
Whether you ascribe to the transformative power of LLMs or hold out for the advent of AGI (Artificial General Intelligence), AI already has the power and the tools to redefine the educational landscape, redefine attainment, and put an end to teaching as we know it. It’s simply a question of putting the pieces together, and scaling it for trickle down.
Everything AI needs to take over education as we know it already exists.
Phase 1 - The Box
Our Guide by the Side
Teachers are not early adopters. For an industry that is, at its heart, liberal, teachers are notoriously conservative when it comes to change.
Educationalists around the world are already writing lesson plans, reports, and examinations using AI tools, but many, if not most, teachers still believe that AI is an issue when it comes to student use, with many viewing AI as an opportunity to plagiarise, rather than a tool to redefine the uptake of knowledge.
Perhaps we need to look beyond the classroom when it comes to the adoption of AI.
Perhaps we need to be radical, and look to the students themselves for our vision of the future.
One of the key debates that happens between curriculum creators is the dividing line between collaboration and collusion. Raising your hand in class is collaboration. Asking your teacher for feedback, or doing your homework with a friend, is collaboration. Getting your friend to fix your maths for you - collusion. So far, so simple, but that gets a bit more tricky when we stop looking at people, and start looking at study aids.
Reading beyond the textbook is encouraged, of course, but if that information was pertinent to top grades, it should have been in the textbook in the first place. Peer study, flipped classrooms, student teachers… these are all considered acceptable learning aids. No one claims that the child with a tutor for an Oxbridge exam is colluding, as long as that tutor isn’t writing the answers for them. But that area starts to get grey when it comes to machine support for student learning.
As I write this I can spy a little blue ribbon six lines up telling me I might want to rethink a comma, but no one thinks - at least not anymore - that I am colluding with Google to make my article look better. Duolingo is a perfectly acceptable approach when it comes to learning a language you’ll later be tested on in school, but Grammarly is contentious when it comes to English teachers. Do we feel the same about a student studying with a learned parent or sibling as we do with an AI? And yet, which one do we think is better at helping: my sixteen year old hormone-driven brother, or an artificial intelligence with the repository of all global knowledge at its disposal? Which do we think has a higher likelihood of getting the answer wrong, especially if his girlfriend is out with her friends and hasn’t texted?
As a long-time teacher of Shakespeare, I am prone to get a scene wrong here or there, and I can never really remember which Henry contains the line “we have heard the chimes at midnight.” I remember Cordelia, but what were the names of the other sisters in King Lear? I quite often get my facts mixed up, but technology doesn’t, not with enough rapidity that we can’t tell when something’s up.
The fact is that the Box has been with us for decades.
Computers in classrooms are not new innovations. Wikipedia has some helpful overviews. Iterative testing is a fundamental part of most countries' curriculum. Google Scholar may be the first repository of all researchers, but certainly not the last of its type. Websites that mark and highlight areas of growth and error are a daily part of every child’s education, if not in their classroom. And no one thinks Udemy coding courses are cheating when you’re studying for your computer science exam.
But the Box is getting smarter. Smart enough to individualise education in a way educationalists have been asking teachers to do for decades.
Technology isn’t supplementing teaching anymore. It’s teaching! And it’s only going to get better.
Stephen Wolfram sums up the ability of the Box nicely when he speaks about the fact that AI is moving from supplying us with information, to identifying the gaps in our knowledge, before then supplementing and reinforcing these in a way no classroom teacher could ever do. And who could deny the benefits in this.
The next few years, Wolfram posits, will see AI training us, not simply with the knowledge we lack, but training us to be in a state where we can fully accept that knowledge, building not just the framework for the knowledge to co-exist in our minds, but the cognitive and meta-affective capability to take it on board in the first place.
In other words, it won’t just be teaching us. It will be conditioning us to learn.
In 2020, there were over 1,000 AI-powered educational apps available on the market. By 2025, AI in education is expected to be worth more than eighteen billion dollars - and that was before the advent of ChatGPT. Alison King’s Guide on the Side is being readily realised as we speak, by a Box, rather than an underpaid graduate who is worried about student loans.
The race to provide that Box is already well underway. Google, Microsoft, Apple, Khan Academy, and a number of forward thinking start-ups are already jockeying to take position in a readily crowded market of multi-capable systems. Artificial Intelligence in Education Technologies (AIED) provide predictive models that can identify at risk students and create personalised learning experiences in a way that only the best teachers can do, and even they can’t do it twenty-four/seven.
SAMR (substitution, augmentation, modification, and redefinition) may have begun as a framework in showing how technology can aid learning, but it’s also an apt model for the impact of AI on Education over the coming years. Whether we like it or not, student essays, their test results, and, indeed, their learning, is already deeply within the collaboration, if not collusion, phase with artificial intelligence. It’s not difficult to imagine that within a year or two, every high school student will have a Box next to them, recording lessons and regurgitating key takeaways, explaining concepts and theories in the cognitively approachable ways that their teacher couldn’t, and preparing them effectively for exams in ways only the world’s best tutors could do so.
What use then for King’s Sage on the Stage? What use for the teacher who knows everything about the subject, when the Box knows as much - and more - as there is to know, including the paper on the topic that was published last week?
What use for an innovative educator trying to reach their students, when every student can be reached by the Box in a tenth of the time.
Forget the S, A and M, in SAMR. Redefinition is a matter of years away.