Following on from previous talks in the seminar series I’m convening at the Open University, my colleague Wayne Holmes gave the last one on artificial intelligence. I found AI perhaps the most thorny of technologies in educational technology. Firstly, how much of it is hype? Secondly, what can it do usefully and realistically in education? And thirdly, perhaps most significantly, what are the potential negatives?
These three questions are at the core of ed tech in 2018, and nowhere are they more prevalent than in AI. So, I was particularly grateful for someone who knows what they’re talking about to guide us through these. On the first of these, Wayne pointed out the prevalence of AI that is already in so much of everyday life. It’s not a future technology he argues, but over the past decade has arrived, from Siri to games, but we often don’t think of these as AI. He highlights the distinction between ‘general AI’ and narrow AI – the former is being human like across tasks and is a long way off, whereas the latter is targeted at a specific function and is prevalent. I think this distinction is important for discerning hype and reality.
Wayne gave a good overview of existing AI systems that are currently in use in education, from intelligent tutoring systems to chat bots. He ended by addressing the risks, including embedded racism in ‘intelligent’ law enforcement tools and Facebook’s move into personalised learning. The role of ethics in AI is now essential, and he highlights the intersection of education, big data and algorithms, each of which has their own set of ethical considerations, and also a particular set when we combine them. It is this kind of analysis that is missing from so much of computing education, tech start up coverage and and ed tech discussion.
Anyway, his excellent presentation is below: