Supporting Classroom Learning with GenAI: The Role of Prompt Engineering

Supporting Classroom Learning with GenAI: The Role of Prompt Engineering

When thinking about using AI in the classroom, a common question on most teachers’ minds is, will GenAI let students jump straight to the answers and skip all the learning in between? No one wants AI to replace real learning or weaken students’ problem-solving skills. Naturally, schools and teachers want to ensure AI will help learning rather than undermine it. So, how do we create a tool that supports learning with structured guidance, not just the answers?

Educational software designers are harnessing AI technology for good, using thoughtfully designed prompt engineering. The outcome? jacTUTOR. We have crafted a purpose-built GenAI Tutor tool that can provide students with personalised support within the context of the curriculum – all thanks to careful prompt engineering.

What is Prompt Engineering?

In simple terms, prompt engineering is the process of creating a set of instructions for a GenAI tool to follow. GenAI is short for Generative AI, and it does just that: generates a response based on whatever question or prompt is put in, drawing on intelligence it has acquired from sources. We can use prompt engineering to formulate a string of multiple prompts to effectively produce something that meets specific criteria or objectives.

When thinking about GenAI tools, top-of-mind for most people is ChatGPT. If you’ve ever used ChatGPT, you’ll know you can ask it questions and it will give you a response. From there, you can feed it further instructions to refine the output, working towards your desired result. The conversation with a GenAI grows as it learns what you want it to produce.

Prompt Engineering sets the rules for GenAI in education.

When it comes to learning, prompt engineering is used to direct the GenAI tool on how to respond to students within the context of that question, lesson, and curriculum. When setting the rules, software developers can define what type of information is given to a student, how it is expressed, and how fast or slow it is drip-fed, so as not to provide all the answers at once. This is what sets it apart from generic tools like ChatGPT.

A purpose-built GenAI tutor offers students relevant and accurate guidance. To give you an idea of why this is important, let’s explore an example for Mathematics. There are many methods of learning Mathematics: a generic tool might draw on many sources, not necessarily from Australia, and use different methods each time. These methods might be different from how the student has been taught that concept in the classroom, causing more confusion than assistance.

Just imagine the benefits for students both inside and outside the classroom, especially during self-study. This technology will truly extend a teacher’s reach and give students 24/7 access to support. Like having a friendly chat with a teacher, when a student is stumped by a problem, they can ask a GenAI tutor a series of questions to point them in the right direction, precisely when it’s needed.

The ‘guardrails’ for relevant learning.

Our very own Claire Lord, Head of Technology, sat down with us to reflect on the exciting work her team has been engaged in with our GenAI tool, jacTUTOR. Lord explains, “Prompts act as guardrails. Multiple layers of these guardrails guide jacTUTOR to facilitate best practice learning within the bounds of the curriculum.”

Providing guidance aligned with learning is key. As Claire Lord further explains, “Through prompt engineering, a GenAI tutor is effectively told how to interact with students and how to facilitate learning that is aligned to the Australian-specific curricula. We are simulating a prescribed educational context that suits the classroom.” So just like a real-life tutor, a GenAI tutor is guided by curriculum and theory, as well as a student’s age and learning stage.

Lord continues, “Its mission is to help students get unstuck. Conversational-style guidance aims to mirror what a teacher or tutor might do or say when a student is stuck.” jacTUTOR is an approachable learning buddy, offering sequential tips to steer the student forward.

GenAI is a tool for differentiated learning.

The beauty of this technology is the opportunities it presents for more personalised learning. Students receive guidance specific to what stage they’re at when solving a problem. This means students stuck at the very start of a problem can get help understanding what the question is asking them. Whereas students who have gotten a bit further and are stuck part-way through the problem can get some ‘next step’ advice.

GenAI can be a wonderful solution to support students no matter where they are in their learning. “jacTUTOR is going to help teachers apply differentiated learning in their classrooms. Teachers continue to be in control of the learning environment, while supporting their students’ differing abilities and empowering students to progress at their pace” says Lord.

It’s an exciting road ahead, for both teachers and resource developers. The team at Jacaranda are helping create guardrails for safe and effective classroom use, so together we can unlock the full potential of GenAI in education.