Knowledge Collapse – Michael Harris in Boston Review

18 hours ago

Mathematical research is described as a free creative art and a form of unalienated labor, driven by the pleasure of understanding.
AI has long targeted mathematics as a challenge, with recent advances leading to predictions that human mathematical power will be outstripped.
The mathematical community is divided between boomers (who embrace AI's potential) and doomers (who fear human obsolescence).
Formalization and proof-checking technologies, such as Lean, have gained traction, but they raise concerns about separating proof from human understanding.
Mathematical understanding is a core value, often described as intuitive and not reducible to formal reasoning, yet AI prioritizes mechanized reasoning.
Historical examples, like the Four Color Theorem and Kepler's conjecture, illustrate tensions between computer-assisted proofs and human comprehension.
The Riemann Hypothesis symbolizes both a milestone for AI and a potential loss for human understanding if solved by machines without insight.
Industry-driven AI projects risk alienating labor, commodifying expertise, and collapsing knowledge by prioritizing commercial goals over shared culture.
Initiatives like the Leiden Declaration advocate for ethical AI use in mathematics, emphasizing autonomy, understanding, and resistance to harmful applications.
Small-scale, non-commercial AI projects show potential to enhance human understanding while preserving mathematics as a gift economy.

Hasty Briefsbeta