Tipping the Scales of Tradition: The Status Quo Bias in Education.
The adoption of generative AI tools among educators is a mixed bag with some forging ahead of others who continue with a business-as-usual approach. Here, I explore why BAU is compelling but limiting.
“Raise your hand if you’ve used ChatGPT”
This week, in a lecture theatre full of educators from a range of Australian schools, almost every hand was up. The speaker, Associate Professor Jason Lodge, then instructed, “Keep your hand raised if you’ve spent more than one hour using it”. At this point, about half of the hands fell, and by the time ‘more than 10 hours’ was posed, almost no hands remained.
To date, the topic of AI in education has focused largely on how students are using tools like ChatGPT and similar Large Language Models (LLMs). By now, most educators have a good sense of why this is worth worrying about; students can use these tools to convincingly do their homework and assignments for them. Naturally, this poses a serious threat to the integrity of traditional assessment tools we use to gauge learning outcomes. However, to respond to this threat in a meaningful way (sorry, banning the technology ain’t it), we must integrate its use into learning activities and assessments. This of course means that educators must be able to use generative AI tools well themselves. So why aren’t they?
The Status Quo Bias
A well-documented effect in psychology is that people prefer to stick with what they’re doing rather than change to something new—the status quo bias. Daniel Kahneman and his colleagues explain that this phenomenon is an example of the larger effect of ‘loss aversion’, where the potential losses from changing a situation loom larger in people's minds than the potential gains. The status quo bias is essentially a comfort zone where the known and familiar are favoured over unknown alternatives, partly due to the fear of potential losses. But what do we mean by losses?
Many of the studies on loss aversion and the status quo bias have been conducted using real-life and fictitious scenarios about economic decision-making. Money is an easy metric to use when thinking about gaining and losing something. However when it comes to teaching, the precious currency prized by educators is time. Between hours of lesson planning, heavy teaching loads, mountains of marking and a declining workforce, it’s no wonder that time-poor teachers are reporting high-rates of burnout. It’s therefore completely understandable that the prospect of spending any amount of their precious time on learning how to use a new tool would be perceived as a loss; the risks of doing something other than responding directly to the demands of the day are tangible and they loom large.
Tipping the Scales
The flip-side of the risk coin is opportunity. Unfortunately, the status quo bias shifts our attention away from potential opportunities, so please allow me now to re-direct your attention, because when it comes to using generative AI for teaching, the opportunities are enormous.
The first and most important thing educators stand to gain by learning how to use tools like ChatGPT is, you guessed it, time! But critically, the trade-off between the time that you spend learning and the time that you end up saving down the line is non-linear. It isn’t the case that if you spend 1 hour now, you’ll save one hour later. It’s impossible to quantify it precisely, but let’s take a rough stab using lesson planning as an example. Recently, teachers trialling a generative AI lesson planning tool reduced daily lesson planning time from 60-90 minutes down to just 60-90 seconds. If it takes, say, 1 hour to learn how to use the tool this week and you had 10 lessons to plan for next week, then you’ve just bought yourself about 10 hours of time. And of course, the relative gains only increase with each week that you apply your learning; after 10 weeks, your 1 hour of learning has potentially bought you 100 hours!
Another fantastic potential gain is creativity. Some of the most impactful tools in a teacher’s toolkit are examples, metaphors and analogies. Explaining concepts in ways that students can grasp them is not an easy thing to do, so educators must skilfully devise a range of examples that breathe life into abstract ideas. And the best kinds of examples are the ones that students can relate to, which requires a huge effort from teachers who must attempt to put themselves in the shoes of their students. Luckily, this is a task that ChatGPT is great at! For instance, check out this transcript where I asked ChatGPT to generate some examples of entropy for year 10 students. While you may not like every example that ChatGPT gives you, the beauty is that you can instantly generate as many examples as you like and select the very best—a process that would generally require a lot of time and effort.
On a related note, by virtue of having access to this wealth of creativity, the opportunities for tailoring learning to particular classes and even individuals are profound. By prompting ChatGPT with a few key details about your intended audience, you can turn otherwise generic class activities and even assessments, into highly refined tools that target the strengths and weaknesses of your students. You might even feed the work of your students into your chat window and ask ChatGPT to generate its own thoughts about areas that need more focus!
Getting Started
Taking the first step in learning something new is always the hardest part, but you don’t need to go in blind. By now, there are numerous guides and YouTube videos targeted towards teachers that show you a wide range of tasks you can do with ChatGPT and other generative AI tools. If you find yourself stuck at any point, I’d suggest searching “How to…[insert task]” into Google and finding a video that walks you through the solution—this simple process has saved me more times than I can count! Ultimately though, there is really no substitute for good old fashioned practice! It’s only when you begin to try things out for yourself, come up against roadblocks and look for ways to overcome them, that you start to develop the kind of expertise and flexibility that will really transform your workflow. Here are some ideas for things to try out when you get started:
Write an email to parents….or even, write different versions of the same email tailored for particular kinds of students.
Plan a lesson
Generate ideas for examples and analogies
Design an interactive class activity
Design some materials to accompany the activity (within ChatGPT4, the paid version, you can generate stunningly good images for this)
Generate a new assessment
Create the assessment briefing
Design the marking criteria
Generate exam questions including multiple-choice, short answer and essay prompts.
Feed in student work and your marking criteria and ask ChatGPT to provide written feedback on the work.
Create tailored instruction and feedback for students
Hearts and M-AI-nds
The above list is by no means exhaustive, but hopefully it sparks some inspiration and maybe even helps to tip the scales of the pesky status quo bias. As we approach the beginning of a new school year where the AI ban will be lifted, it’s vital that educators know how to use this technology, for the sake of their students, but also for themselves. While embracing AI in education offers efficiencies and creative opportunities, it's important to remember that these tools are not here to replace teachers. Instead, by using AI well, educators can redirect their invaluable time and energy towards the most human aspects of teaching: building deeper connections, understanding, and engagement with their students. This is where the true heart of teaching lies.