Game Over for Maths A-level
The combination of ChatGPT with its Wolfram plug-in just scored 96% in a UK Maths A-level paper, the exam taken at the end of school, as a crucial metric for university entrance. (That compares to 43% for ChatGPT alone).
If this doesn’t shock you, it should. Maths A-level (like its equivalent in many other countries) is held up as the required and essential qualification for much of our populations—the way to be prepared for our upcoming AI age. And yet, here it is, done by those very AIs, better than most of our students.
Wrong conclusion: ban it. Right conclusion: change what humans are learning so they step up a level, and don’t compete with what AIs do well.
While the combo of ChatGPT and Wolfram starkly makes this point, the problem is not new. It’s simply been dramatically exposed.
You might say, but ChatGPT can write good history essays too, so why are you singling out maths A-level? Because it’s been clear for over a decade that the subject-matter is dramatically wrong, not only the assessment technique, because computers pre-ChatGPT were already central to real-world maths, but not eduational maths.
In say, history, this has not been the case. You’d type, not write, you’d access information differently, but the subject has evolved, not been turned on its head by computers, at least up to now. The real-world change from ChatGPT for history is yet to manifest itself, so our target for change is not yet clear. But change there will need to be.
What is clearly unfortunate in all subjects, is how assessments have become more and more formulaic, while what’s needed from the human (the human value-add) is doing what the computer doesn’t—becoming more creative, less rote; having knowledge of what to do, not just facts to know or microprocesses to follow. Formulaic assessments are driven by ease-of-grading, so a simplistic grade can represent years of human achievement; the whole process is itself a failure of computational literacy in our society.
Back to maths. When will governments, universities, the educational establishment wake up?
Stop funding millions to “improving teaching of maths”, when it’s the wrong subject, whilst failing to fund a fundamentally new computer-based curriculum, one to achieve computational literacy for all. Also wrong is education-establishment, inward-looking curriculum setters, relying on a pre-AI past to set the future. Will ChatGPT finally be education’s wake-up call, or yet another failure of political leadership to make critical change?
The UK government’s current policy is “no change of curriculum before the next election” so we don’t upset teachers and schools. The opposition’s top education policies have no mention of “computational literacy”, or indeed any fundamental subject change in sight. Both totally miss the point—like deciding we’re going to make incremental changes to the cavalry for fighting wars, not to upset the army, when tanks, jets and drones are operational.
And I use the UK as an example here. Almost every other jurisdiction is in denial.
ChatGPT’s success on maths A-level needs to be the trigger for long overdue change. As I wrote 3 or so years ago in The Math(s) Fix:
This is game over for the traditional maths A-level as a core subject. When that will happen, how the change will unfold, the depth to which we’ll persist in burying our heads in the sand before it does—and crucially which countries will win and lose, is hard to know. But make no mistake, one day this will be over—and not a moment too soon.