AI and the Future of Mathematics
Main argument
The speaker argues that mathematics remains valuable—and may be especially worth pursuing now, even if AI becomes better than humans at proving theorems or eventually at every part of mathematical research.
AI may force changes in how mathematicians work and receive academic credit, but it does not automatically make mathematics, mathematical education, or human engagement with math obsolete.
What mathematicians do
The video describes several connected activities:
- Prove theorems
- Choose worthwhile questions and make conjectures
- Build theories that provide broader ways to understand mathematics
- Apply mathematics to science and engineering
- Communicate ideas clearly, in writing and through talks
The speaker sees theory-building—not proof production alone—as the deeper aim of research.
Why mathematics is valuable
- Applications: Pure mathematics can lead to practical uses, sometimes many years later. Examples include complex numbers, non-Euclidean geometry, and number theory contributing to cryptography.
- Transferable skills: Mathematical training develops problem-solving abilities used in fields such as computer science, finance, machine learning, and AI safety.
- Aesthetic enjoyment: Many people value the satisfaction of solving problems and the new perspectives that mathematical ideas provide.
Where math’s funding and broader benefits come from
Mathematics is supported by university teaching, public grants, private foundations, and industry. Teaching is a particularly important funding source. Math departments train many students who later work outside mathematics, spreading problem-solving skills through the wider economy.
The speaker also emphasizes math’s cultural and educational value: people enjoy learning about mathematics, and teaching is a social, motivational activity—not merely the delivery of explanations.
If AI surpasses mathematicians
The speaker considers increasingly ambitious possibilities:
- AI proves theorems: Proofs may become more like raw material for understanding. Researchers could focus more on interpreting results, connecting ideas, and building theories.
- AI chooses questions too: Humans may still contribute by making sense of the results, developing theory, and applying it.
- AI builds theories: Human understanding and theory may matter more for connecting mathematics to real-world problems.
- AI handles applications as well: Funding motivated solely by applications could shift toward machines. However, human education, transferable skills, and the enjoyment of doing math could still justify supporting mathematical institutions.
Across these scenarios, the value of practicing mathematics remains comparable to exercise: a machine’s greater ability does not remove the benefit people gain from doing it themselves.
What may need to change
- Academic credit: If AI can generate proofs, proof production becomes a weaker measure of human contribution. The community may need to recognize theory-building, insight, exposition, and other forms of intellectual work.
- Training and assessment: Traditional problem sets may need rethinking as AI changes how students solve problems.
- Communication: Oral presentations and human explanation may become more important for showing understanding and helping others care about ideas.
- Community: The speaker encourages mathematicians to discuss what they value in the subject, rather than confronting these changes in isolation.
Advice for students and researchers
The speaker’s practical recommendations are to:
- Reflect on which parts of mathematics you genuinely love: proving, theory-building, explaining, applying, or pursuing human-led mathematics.
- Consider how AI fits with your own goals, rather than making decisions solely out of fear.
- Give talks regularly and develop strong presentation skills.
- Help shape what mathematics and academic contribution should look like in the coming years.
- Tomorrow: take a long walk with another student and talk about why you chose math and what you want your work to contribute.
The speaker’s central encouragement is that the mathematical community is unusually open to new directions right now. That uncertainty creates an opportunity for new researchers to help shape the field.
