Thoughts on Education for an AI World

The Social Context of Education

The complex relationship between our collective and individual interests is evident in the way we frame the purpose of education and how we balance its competing objectives.

Interest-driven learning makes a lot of sense to most people, but how do we assess whether pupils are genuinely engaged and interested rather than merely being compliant? And how do we ascertain whether, by pursuing their interests, children and young adults cover the material their societies think necessary and worthwhile?

We sometimes imagine that our learning-journey is like a path that stretches away in front of us and leads towards a fixed destination, but what if there is no path and even less a fixed destination? What if knowledge is like a fractal that reveals more and more of itself and - far more significantly - creates more of itself at every stage?

In an age of artificial intelligence it matters more than ever that we get education right. But education arises from a cultural-social-political-economic system in which it is so deeply rooted that it proves almost impossible to reform it before that system reforms. In the UK alone over 9 million children are educated by over 400,000 teachers, each of them with different backgrounds, skills, interests, strengths and weaknesses. So it is not surprising that education’s first knee-jerk response to AI has been to entrench, double-down and pretend that if we ignore it, it will go away. It will not go away.

In the first of these posts we explored the way education has tended to default to understanding itself in terms of content-delivery, specifically the transfer of information from the teacher to the pupil variously aided and abetted by text-books, libraries and latterly the Internet. The particular way this mirrors early attempts program machines to act intelligently is that both assume that what matter are knowledge and deliberate, painstaking application of predetermined algorithms.

In a very short piece written in 2019 called The Bitter Lesson, Rich Sutton of the University of Texas at Austin made the observation that human intuitions about how best to program computers to solve problems and play games have been shown to be flawed; in particular, the approach that believes that what we have to do is reproduce in machines - specifically in different versions of AI embodied in Large Language Models (LLMs) - the same thinking-processes that we believe human beings to engage in.

This marks a fresh start for the EDSIL blog, now hosted on our new platform.