AI in mathematics may be less like a calculator and more like a tractor

What is happening to mathematics as AI gets better? Perhaps it helps to compare this moment with earlier changes in how people work and how things get done.

The first analogy is mechanized agriculture. A tractor can do the work of many people, but it still takes people to operate. For a while, a small number of farmers will steer machines far stronger and more tireless than any person. AI may play a similar role in mathematics. Over time, more of the steering may be automated too, and an entire profession is gone!

A second analogy is the telescope. It did not replace astronomers; it changed what they could see. AI may become a kind of intellectual telescope for mathematics, bringing distant connections or unexpected conjectures into view. A mathematician’s job then includes deciding which glimpses are real, which are artifacts, and which deserve a closer look.

A third analogy is the computer. Computers did not make mathematics disappear; they made different mathematics possible. AI could push that boundary further, not only carrying out calculations but proposing approaches and assembling proofs.

These analogies are incomplete and, emphatically, far-cries given that AI is poised to completely take over. A tractor does not invent its own goals, a telescope does not suggest what it sees, and a traditional computer does not usually choose its own strategy. AI can do all of these things, which is why this transition feels different.