Mathematician Po-Shen Loh argues that faster AI proof production leaves a human responsibility: deciding which research directions serve people. His case for maintaining a mathematical community rests on that responsibility, rather than on a prediction that humans will always outperform machines.
In his argument, informed oversight requires active expertise, so researchers must keep practising and train successors. He also proposes giving teaching and other work benefiting people greater weight in academic hiring and promotion. This is a proposal for how the profession should adapt its priorities as its tools change.
