Lately, I've been diving deep into a sea of articles expressing a collective concern over AI's role in schools. The common theme? A profound worry about AI's knack for churning out essays that students could submit as their own.
Reflecting on my journey through the eras of computers, the internet, and now AI, I've realised something important: this fear of technological advancement in education isn't new. For the past 40 years, we've been continuously adapting to technology's evolving role in learning.
Let's take a nostalgic trip through the past four decades, exploring the creative ways students have sidestepped homework and the lessons we've learned from them.
Library books will ruin education
In the pre-internet, pre-multimedia era the de facto way to do homework was to pop to the library, research your content from a bunch of different sources then distil this information into a coherent and well reasoned piece of writing that could be submitted to the joy and delight of your teacher.
Except that's not what happened, dozens of children would all go to their local library over the course of the week and laboriously copy the information in the first encyclopedia they found out by hand - verbatim - and hand that in as research.
Now some research backed pedagogy suggests that this is actually quite a powerful tool for information retention, it's passed through the phonological loop of the student at least once so the next time this topic is discussed in class it encodes in memory better and long term retrieval of this information is improved. It actually becomes the foundations of retrieval practice.
Jokes on us - from Millennials back - when we thought we were getting one up on the teachers we were actually priming our brains for better retention of the topic in class 😅
Copying each other's work will ruin education
Join me here if you've ever rapidly copied a friend's homework during a break-time before a lesson 👋 yup - me too folks.
There's an art to copying homework, get the main points down, change it just enough so that it reads differently and you don't get picked up by the teacher - this is a tale as old as time and probably has roots in Socrates himself.
Okay, so at broader level academic plagiarisation is a real problem - and no one wants to ever find that on final coursework submission two students work are too similar - but on a simpler, day-to-day level some of the research suggests that high emotional response, such as panic, improves the encodinging of memory. Just like the scenario in the library example, students panic copying and transliterating other people's work is actually quite an involved thought process that results in solid information retention when the concepts are revisited.
Another instance where thinking we were getting one up on the teachers was actually part of the expectation of developing an understanding. That terror of getting caught, and the meta learning aspect of converting someone else's understanding into your own words has significant benefits for students if they're doing it well.
The Age of Encarta
Here's where things started getting a little dicey - the problems before this have been problems for hundreds of years, but as technology took a drastic leap and suddenly we have every student with a computer and access to a CD-ROM encyclopedia that contains a shed load of information about nearly everything they could be asked about (and an epic game to play to take up those rainy afternoons too!).
So - teachers were relying on the library method of research but what electronic encyclopedias allowed was for students to hit print and get a copy of the text they could just hand in 😱. Teachers got wise and insisted that they wouldn't take print outs from the software itself - and that's the reason why an entire generation learned to use copy-paste and changed the fonts.
Unfortunately there are no secret pedagogical wins for this, students would produce work without reading it, there was no summarisation, no condensing of an article and it never hit that phonological loop. Teachers had to start changing up the way they set this kind of research homework - "show two sources and summarise" was common, as was simply giving zero to any direct copies from Encarta.
That skill of summarisation ensured that the reason for this kind of work, the pedagogical principle behind pre-learning, was maintained - it was just much more likely that students wouldn't have read the initial material and so lessons started relying less on students having definitely done the pre-reading and spent more time including in-lesson activities to work around that.
Alta Vista 😅👵
The early days of the internet meant that it was much harder for teachers to identify the sources that students were copy and pasting from, the - select two, and summarise - method started to fall apart when there were 100 articles out there about the same thing and students could track down a page you'd never seen before. This changed the game again, and I even managed to score an A+++ from a teacher for a homework I'd submitted and never actually read 🫢 apparently leaving the contact details of one of the organisations I'd copied and pasted from was actually deemed to be good research back in the day…
The problem was that old-style search engines were difficult to use to find results that were relevant, and even then it took some cognitive overhead on the part of the learner to sift through the data. In many respects this model followed the Library model of research above and although students didn't often read all of what they were submitting they did have to engage their brains, their phonological loops and critical thinking skills in order to find the most relevant results to get copied and pasted in the first place.
This is one of those weird technological changes where the pedagogical benefits moved back to a previous state and after a period of adjustment the benefits of the task were still being felt.
Google and Wikipedia
The advent of decent search and a central repository for the knowledge of humanity meant that it was now the work of seconds to type in a search term, find the wikipedia article and copy and paste that content into a word document. Whilst this gave a generation of young people the skill to paste without formatting and removing hyperlinks, it sort of went a little full circle back to those Encarta days of every student submitting the same content for their work. But we know how to deal with that right?
Well actually, there was a really nice hack to teach some meta learning skills that we used to be able to do before those pesky Wikipedia editors got wise to us - we'd set some work and head to the article to contribute the line "if this is in your homework you get zero" to the text, it would usually take a week or so to get taken down, but marking that work was a cinch - highlight the copy and pasted line and write a big zero on the paper and the student always made sure to read the homework before handing it in.
But hey, that's the world we've been living in for fifteen years. Generations of teachers and students have worked in this system where you can't just set a "research this" homework, it's more about a "write an essay that…" lesson that leverages the content they've blindly copied and pasted, it's about adapting the practice of "research" to the modern reality of it and we've worked that way for a while.
AI and "Write an essay about…"
That's where the most recent breakthrough in technology has really started worrying people, where do we stand with cheating and what is the point of a homework in a world where students can simply copy your question into chat gpt and get a thousand word essay back.
We haven't just lost the essay as a way of assessing understanding, but seemingly comprehension questions, well… any questions really. Erk, that sounds like a bit of a problem.
Hold on, have we broken assessment?
Well, no. We just need to adapt. We need to engage in AI, we need to be more open in our use of it so that students are not using this blunt tool in secret and not actually learning anything from it. Just like our forebears for the last 40 years of dealing with cheating at schoolwork, we need to work out what AI can enable in our assessment practices and use it. It's a bigger job this time, but we can get there.
Adapting What, How and Why we Assess
What is the gold standard of academic rigor? Defending a thesis? How about we pivot our assessment strategies to focus on that? Here's a framework to guide us through the integration of AI into our assessment processes:
- Emphasise Critical Thinking: Instead of asking students to write an essay, we could ask them to critique an AI-generated essay. This would require them to use their analytical skills to identify strengths, weaknesses, and gaps in the content. Heck - we could even get them to keep prompting the AI with changes to get it as close to 100% on the mark scheme as possible, giving them a better understanding of the requirements of the essay structure and content.
- Oral Examinations and Discussions: Leveraging the traditional viva or oral defense format can encourage students to articulate their knowledge verbally, demonstrating understanding beyond what an AI could generate. You can even simulate this with AI Tutor bots, by prompting the bot to take up opposing views and argue against the points made by the students you'll develop their ability to construct effective arguments and enhance their knowledge immeasurably.
- Project-Based Learning: Tasks that require students to create something unique, such as a project, presentation, or experiment, provide tangible evidence of their learning that is much harder to fake with AI. AI Tutor bots can be brought in here to help students ideate, plan and execute their projects - even critiquing the end result against rubrics or mark schemes.
- Collaborative Work: Group projects can be designed to ensure individual accountability and focus on interpersonal skills, project management, and collaborative problem solving—skills that AI cannot easily replicate - although AI can be used to augments a students ability to coordinate all of those skills.
- Customised Assessments: Teachers can design assignments that require personal reflection or application to students' own lives, which is something that AI's really good at doing at scale actually!
- Use AI as a Tool, Not a Crutch: Encourage the use of AI to aid initial research, draft outlines, or explore different viewpoints, but always with the expectation that students will add their own voice and critical analysis. Getting students familiar with using AI in the classroom is the step you need towards stopping it being some hidden secret thing they're using, and turns it into the ultimate learning tool. Get your AI out of the darkness and into your classroom.
- Process-Oriented Learning: Shift the focus from the final product to the process of learning, where students document and reflect on their research, thought processes, and learning journey. Another thing that AI can prompt students to do, help organise their thoughts and structure their answer, but importantly can't do for them.
- Open-Book, Open-Web Assessments: Embrace the resources available, including AI, and set assessments that demand high-level thinking, even with all information at hand. Questions should require synthesis, evaluation, and innovation.
- Adaptive Testing: Use technology to create assessments that adapt to each student's level of understanding, providing a more personalized measure of their knowledge and skills. Guess what technology is good for this? Shocking no one, it's AI.
The integration of AI into education is not the end of learning or assessment; it's an opportunity to innovate and redefine our approaches to teaching. By recognising and adapting to the changes AI brings, we can harness its power to enhance education instead of fearing its potential for misuse.
You know what's a really good platform for enabling those superpowers in your classroom? Mindjoy.
But I would say that. Why don't you try it out and find out for yourself?
