Stay informed with free updates
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.
The development of artificial intelligence is one of the biggest changes in how we live and work since at least the invention of the internet. It will change how we all go about our days. It will change, and is already changing, how we are governed.
As a result, there are calls in almost every wealthy country to teach “how to use AI” in schools or to “put AI on the curriculum”, or else to rework teaching and assessment so as to reflect just how much of what we ask school children to do can already be done by AI. In any given week, I receive at least one email a day from one campaign group or another calling for exactly this.
This debate is about a new technology, but it is part of a venerable argument in education policy about the extent to which school curriculums should be “skills-based” or “knowledge-rich”. Should schools be equipping children with the skills they will need in the future or giving them a broad foundation of knowledge?
One reason why “knowledge-rich” curriculums have outperformed “skills-based” programmes is that we are terrible at predicting the future. How could it be otherwise? Someone starting compulsory education in Ohio in 1977 couldn’t possibly have been expected to know that by the time they left school, their state would have experienced significant deindustrialisation, the cold war would have ended and personal computers would have started to become affordable for much of middle America.
What “skills” will today’s children need in the world of AI? History teaches us that those we once thought would ensure a reliable income forever are no guarantee of any such thing.
Coders, who relatively recently were considered to be the inheritors of the earth, now face an uncertain job market and may never go back to the same position and security they enjoyed, say, five years ago. Or, to take another example, in oil-rich Alberta in the middle of the last decade, unemployment among geologists — also previously considered an ironclad vocation — was at 50 per cent.
In some ways, the distinction between skills and knowledge is a false binary. As Daisy Christodoulou, director of education at No More Marking, puts it, it is like asking what “the right balance is between ingredients and cake”.
No More Marking, which uses AI and other tools to reduce the amount of teacher time — the most expensive and also most valuable resource in schools — spent on marking, is itself a good illustration of the difference. Automating marking and other administrative tasks frees teachers up but, to take advantage of the extra time available, they need the knowledge base to understand what they are doing.
Many of my friends who work in skilled trades make the bulk of their income in repair work, but their training didn’t begin with a leaky tap or faulty wiring — it began with acquiring a layer of fundamental knowledge. That’s the other reason why knowledge-rich curriculums are better than skills-heavy ones: it is knowledge that allows you to deploy and use your skills.
The same is true of AI. Part of deploying this technology effectively, in a range of fields, is having the necessary knowledge base to understand how to use it and to make judgments about whether it has given you the right or appropriate answer. As AI becomes more advanced, my suspicion is that our own knowledge will matter more and more, as we will need to be able to justify and explain why we have picked a particular option and why AI has reached its decisions.
That isn’t to say that AI shouldn’t change what is on our curriculums. The curriculum in England is an example of the right approach here, in that it aims to give students the ability to understand the fundamentals of how coding and computers work.
While knowing how to code may not be a useful skill in the workplace of the future, understanding how code works and how it is that computers “know” things is going to be vital. We will be navigating a world where being able to understand why AI behaves as it does will almost certainly be crucial in assessing everything from how we are governed to whether we are receiving adequate service.
A curriculum that equips children to use AI, and to live and work in a world shaped by it, will still be one that imparts knowledge about the technology (and about computers more generally), rather than honing particular skills. We won’t be equipping the next generation to get the most out of the new thinking machines if we aren’t teaching them how to think for themselves first.