AI in Education
Improving student knowledge and understanding of AI
This is the outline of an introduction I gave at the Westminster Education Forum’s ‘Next Steps for AI in Education’ event, Feb 2024.
My perspective on teaching AI in schools comes from being a parent and my role as an AI Scientist. I’m not a professional educator. I have been into schools to talk about AI, but primarily I’ve spent my career building AI models. My role now is as a consultant working with organisations implementing AI technology.
When it comes to the topic of teaching AI in schools, the first thing I think is important to note is that while AI gets characterised as a Computer Science topic, it’s really a very interdisciplinary subject. That’s important in the context of education and how to teach about it. It’s true that there are several key technical foundations underpinning today’s AI — maths, programming and data literacy. But when it comes to applying AI technology, it’s a truly inderdisciplinary effort. AI might be deployed in any domain, Biology or History for example, where subject matter expertise interacts with the technology.
The past decade has seen a huge boom in AI, and the pace only seems to be increasing. There are new products every day, and many new research ideas and papers that get published. Our frontier technology is changing all the time. Schools cannot be expected to keep up with all of these changes. We don’t expect that in any other field, and AI is no different in that regard. Still, AI has been around for a long time, and there are many fundamentals which remain unchanged. Of these, I would argue that data literacy is the most important for schools. When I talk about data literacy, I mean collecting, processing and analysing datasets, and being able to reason about data in the aggregate. Data literacy has been important for many years, but is becoming even more so in the face of AI.
Yet another aspect that we shouldn’t forget is the critical thinking around AI. Ethics and Responsible AI is an important conversation going on right now, and our students are capable of understanding these nuances. Let’s take the example of deepfakes, which are very easy to produce now, and with this being an election year are likely to be on the rise and impacting the world. As individuals we will have to start thinking very carefully about what content we trust, and how we make those decisions. Data literacy plays into this, along with critical thinking about the capabilities and limits of machines.
It’s also worth remembering that today’s students have grown up with AI. Siri was launched in 2011 — that’s 13 years ago. The iPhone was launched in 2007. There isn’t the same ‘wow’ factor about AI and technology for school students as for those of us who grew up before the current boom. In my experience of talking to students, they are excited and eager to learn, but AI is a normal everyday part of their lives that they’re learning about partly by virtue of living with it.
To conclude, I believe that there is an opportunity ahead of us to bring elements of AI into school education. The maths and programming skills that form the basis of AI and ML are advanced, and typically taught at age 16+. It’s important to teach these skills to train the next generation of engineers who will build AI technology in the future. Yet, critical thinking and data literacy can be infused in the curriculum at a much earlier age. The way forward though has to be driven by educators, who are working with technologists, to find the right path for AI education in our schools.
