A keyboard lit from below. The 'enter' key is replaced by a key saying NEWS.

I talked to Jonny Greatrex, Senior Lecturer in Journalism and Media, about his decision to use AI to update a previous learning resource, an interactive game that used H5P to provide work-like experience for students of data journalism.

We worked with Jonny in 2019 to support the development of the original version of the game. Using H5P, a tool which can be used to create interactive digital content, Jonny guided students through a choose-your-own-adventure style game where they saw content in the form of emails from an ‘editor’ and used their data skills to provide accurate information for a story. (See our blog post Embedding Employability Using Online Games.)

While the game has worked well for years, Jonny thought that there were opportunities to use AI – specifically an AI agent, which can use a structure of layered prompts to shape and guide students’ use – to make the game feel less ‘clunky’ and more of an authentic experience that students may expect to have in their future work.

We talked about AI, the things he’s learned, and the advice he’d offer if you want to use it ambitiously yourself!

Beth: Can you tell me a bit about the game?

Jonny: The aim was to use the AI agent tool in our institutional Copilot software to create something for the students where they felt they were interacting with an editor as part of a work experience situation or scenario, and that editor was asking them to carry out a series of pieces of data journalism work.

I wanted it to be similar to a professional context, so that the students can see how these slightly esoteric, slightly intangible skills they’ve learned in Excel will be used in a professional context to generate journalism.

It recreates what a journalist would do, sat at their desk with a computer and Excel and the data. So it’s not like rote learning, it’s not “learn these button presses for no reason whatsoever”, which I think is most people’s experience of Excel at some point or another.

Beth: What made you want to do this with AI?

Jonny: I suppose, if I’m honest, what made me want to try this with AI was to see whether this developing and emerging so-called ‘world-changing’ technology could do things better or easier or more efficiently than the previous iterations to a make it less time consuming for me as a tutor. Could it improve the experience for the students fundamentally to make it more realistic, more lifelike, and so therefore a better learning tool? Could I streamline the workflow for tutors to create this sort of tool for learning with AI?

Beth: So you wanted an improvement in the experience for both parties, essentially. Saving time while continuing to maintain the quality. How did that work for you?

Jonny: I learned a lot about AI, Gen AI, whatever you want to call it, so in that sense… it worked really well.

Because it’s like all technologies, all things you learn, you need to have a go with it, fail, break it, have it break you a little bit. And that is essentially what the learning process is. Unless and until you’re engaged in a problem-solving task, trying to apply something new that you haven’t done that way or in that fashion before, you’re not really learning. You can be listening to something, you can be watching something. But for me, that is what learning is.

So I suppose how did it work for me? It was a fantastic learning experience.

Beth: One of the criticisms and the benefits of AI is that it makes things frictionless, that it reduces that sense of struggling against something, the grit that is essential to the learning process. So it’s really fascinating that you’ve highlighted that grit and struggle as the benefit.

Jonny: AI is fantastic when you know what you’re doing with it. As in, when you’re already an expert in something, it’s fantastic in helping you enhance what you’re doing.

You need the ability to recognise that the perception, the perfection is a facade, essentially. The ability to sort of dig deeper and say, actually, it’s not quite doing what I think it is, or it’s not quite doing what I want it to.

Beth: So what advice would you give for anyone who’s looking to experiment with generative AI? What kind of things did you keep in mind as you were doing it?

Jonny: How do you learn anything? You know, you do it, you have a go.

I think my advice would be to isolate things that it could possibly do a bit better for you.

What I mean by that, is that getting it to replace a human with a deep level of expertise accrued over 10 years of coding, for example, it’s clearly stupid in a sense to have it do that, because basically it’s a very good guesser. But from trying to use it for this game, which was a fairly grand thing in some senses, I identified a couple of other ways I could use it.

So my advice is have a go with AI, because when you don’t quite get something right, you can be like, oh, well, that wouldn’t work, it’s too complicated. But I bet it can do a really good job on this.

Naturally, this is not the end of the story for Jonny’s work with the game. Next up he will be working to make it more flexible if students provide a partially correct response, and he is aiming to trial it again with undergraduate students this term. We wish him all the best and hope to be able to share more about how it develops in future!

Author: Bethany Witham

Learning Technologist

SOC Digital Learning Team
School of Social Sciences
Nottingham Trent University