Quite rightly, therefore, AI is being seen through sceptical or examining eyes. How should we, as universities, respond? Some of the sector has responded by building walls, with detection software, bans and stricter policing. Other institutions have bought off-the-shelf solutions. To many it is seen as a threat to academia.

We’ve been working with the team at City St George’s, University of London, alongside a wider group of academics, to put that assumption to the test. The group includes Baba Sheba, Professor of Digital Education and Transformation at City St George’s and the project’s academic lead; technologist and neuroscientist Timo Hannay, and philosopher Tom Chatfield, author of Critical Thinking. What we found is that AI is only a threat if you hold on to old methods of learning and assessment. If an essay can be generated in seconds, it’s rendered a fragile proof of understanding.

So what if you flip the script? What if, instead of keeping AI out, you let it in to serve the very thing it’s supposed to threaten? That question led to Duologue, a platform that uses AI to actively develop students’ critical thinking. At Bond & Coyne we’d call that starting with purpose, not product, and it’s advice we give clients all the time.

It also comes down to “who decides”. The distinction we kept coming back to was edtech versus pedtech. Edtech is built to scale, to sell and to satisfy a procurement committee. Pedtech starts with what learning needs and works backwards. Much of what’s being sold as the second is really the first in disguise. City St George’s decided the people shaping AI in education should be the ones doing the teaching and learning, not the ones selling the software.

Duologue works the other way round from much of the AI we see out in the world. Instead of answering your questions, it asks them, using Socratic dialogue to push the conversation forward while the learner does the reasoning. Because that thinking happens in the open, the learning process becomes the assessment. It tracks critical thinking as it develops, rather than judging what a student can memorise or write down at the finish line.

It launched with three modules (critical thinking, AI literacy and ethics, and computational thinking), developed with partners including Sage Publishing, the Raspberry Pi Foundation and Wolfram Research Europe.

As development and brand partner within the Duologue collective, Bond & Coyne’s focus has been to create a university-ready experience, with the potential for a commercial product that reaches beyond academia. In many ways, this was both an academic and design/UX experiment. Our aim was to deliver a pedagogically-sound experience that supported personalised, human interactions. And it turned out to be surprisingly social. Take Duologue’s “Labs” – small learning challenges that started almost as a side project. Now staff say they can’t stop until they’ve “beaten” them, and students crowd round a single screen in groups of three and four. When the AI handles the structure, people get more room for the conversations only humans can have.

This doesn’t mean every university should build its own AI. Asking institution after institution to fund its own tools would be unrealistic and hugely expensive, especially with budgets as stretched as they are. Instead, we’re in talks with other universities about pilot schemes, and the interest goes beyond higher education to government organisations, healthcare bodies and schools. The ambition is to share what’s been learned, so others can shape it to their own needs rather than starting from scratch. Real collaboration across the sector is the dream.

Many projects like this start with a hypothesis, question or an idea that’s brewing within an organisation. The brief does not have to be ready in order for exploration to begin – run in the true spirit of ‘university’, incredible things can happen. 

Let us know if you’re considering how AI turns up in your organisation and let’s see if it sparks a project that delivers the benchmark for your sector.