On October 31, I joined an online seminar titled Mathematics in the Age of AI, delivered by Dr. Hamza Waseem (Syracuse University, London, and the Mathematical Institute, University of Oxford) for the National STEM School in Pakistan. The audience was students aged 15 to 18. Teenagers are rarely invited to such conversations, even though, as Dr. Waseem observed afterward, “they are the very ones who will inherit a completely new mathematical ecosystem by the time they reach university.” Over one afternoon, we explored what mathematics really is, how and why we do it, and what schools often get wrong about teaching it. We also weighed the looming prospects and dangers of AI. Dr. Waseem called the session “cathartic,” and I left feeling the same. I also left with a renewed conviction that mathematics is a profoundly human activity, a shared heritage that we, as future teachers, are responsible for protecting.
What mathematics really is
Following David Bessis’s Mathematica, Dr. Waseem drew mathematics as a loop. On one side is official maths (logic): proofs, axioms, calculation. On the other is secret maths (intuition): dreaming, feeling, guessing. Each feeds the other. Grothendieck could not read a mathematical text unless it aroused “mental images, intuitions that will give it life,” and Einstein said conventional words came only “in a secondary stage.” Bill Thurston’s classic essay On Proof and Progress in Mathematics (1994), which heads the earlier readings on Dr. Waseem’s reading list, asked not how mathematicians prove theorems but how they advance human understanding. Reuben Hersh’s 18 Unconventional Essays on the Nature of Mathematics goes further: mathematics is a human, social and cultural activity, not a set of truths floating somewhere outside us. On this view, a proof is the evidence that understanding happened, and understanding happens in a person.
How and why we do it
We do mathematics by moving around Dr. Waseem’s second loop: problem-solving feeds understanding, and understanding feeds problem-solving. His own journey began in 2008 with a Kangaroo competition certificate urging him to “continue training in problem solving.” The why matters most. G. H. Hardy called the mathematician “a maker of patterns,” like a painter or a poet, and Francis Su, whose essay “Love of Humanity Is Not Enough” is on the list, argues in Mathematics for Human Flourishing that we do mathematics because it meets deep human desires for play, beauty, truth, justice, and love.
What schools often get wrong
Bessis’s essay “The Fall of the Theorem Economy” gave me the sharpest lens. Research mathematics has long rewarded theorems as a proxy for understanding. That proxy worked because no human could prove something without first understanding it, and AI has broken that link. Schools run their own version, an answer economy, in which correct answers and grades stand in for understanding. Students learned long ago that this proxy can be gamed, first with memorized procedures and now with a chatbot. Paul Lockhart’s A Mathematician’s Lament reminds us that mathematics is an art, yet we teach only its official half. If school mathematics means producing answers, the machine has already won. What the machine cannot replace is the student who notices a pattern, makes a bold conjecture, is wrong, and finds out why. Grant Sanderson’s guest essay on Terence Tao’s blog, “If math is more than proof, we need to better celebrate the rest of it,” proposes rewarding great explanations as seriously as great proofs. Classrooms can do the same by grading explanations, not just answers.
The looming prospects and dangers of AI
The list deliberately holds both moods. On the hopeful side are Daniel Litt’s “A beginning for mathematics,” Stefano Gogioso’s “Why I am excited, as opposed to frightened,” and Sanderson’s video “There’s never been a better time to go into math.” An AI can be a tireless sparring partner, example-generator, and proof-checker in Lean. On the other hand, the 722 machine-written papers OpenAI released on October 6 remain largely unverified: three were retracted within a day, and some formal proofs established weaker results than their papers claimed. In September, 25 Fields Medalists, including Tao, Scholze, and Viazovska, endorsed the declaration “A Severe Misalignment of AI in Mathematics (https://mathandai.org/)”, warning that “solving problems is only a tool and proxy for achieving the primary goal of conceptual understanding and insight.” In a classroom, the quieter danger is a student who hands the whole loop to a machine and gets answers without ever being changed by them.
Our shared heritage, and what we must protect
“We participate in a millenary tradition,” writes Amar Hadzihasanovic in one of the short pieces on the list. That tradition runs from Babylonian clay tablets, through Euclid, Brahmagupta’s zero and al-Khwārizmī’s al-jabr, to a student in Surrey sketching a parabola today. It is our common inheritance across cultures. As Henry Wilton warned, authorless papers risk implying that mathematics is “no longer a human endeavour.” Chess offers some hope: computers have beaten world champions for decades, yet more people play chess than ever, because the point was always the human playing. Kevin Buzzard asks whether “to grieve, or not to grieve?” For a teacher, the answer is to teach. Dr. Waseem closed by asking what a student must aspire to, and I take that as my design brief. My answer is understanding. That means estimating and conjecturing before calculating, asking “why?” as often as “what?”, celebrating thoughtful wrong guesses, and using AI to sharpen intuition rather than replace it. The machines may write the theorems; our job is to make sure people still understand them. Progress in Maths means human understanding, not just an answer, because learning has never been only about the answer; it’s the process, it’s the struggle, it’s the mistakes, it’s the moment that it finally clicks, and we finally call, ‘Aha!’ That’s where the love of learning happens.
No comments:
Post a Comment