Headlines and inside stories: understanding and trust in AI for mathematics, science, and engineering

Guest post by Tapio Schneider (Caltech, CliMA). Starting from OpenAI’s Sept 8 forced Navier–Stokes blow-up result, he argues that “theorems are the headlines, proofs are the inside story”: proofs carry understanding, not only truth. In science and engineering, predictions such as climate projections or aircraft designs must often be trusted before they can be checked, so “an auditable causal chain from assumptions and input data to the predicted outcomes is what makes predictions trustworthy”. He proposes keeping AI inside auditable scaffolds (e.g. conservation laws), with learned components such as turbulence closures tested against verifiers, and says humans remain essential for understanding and trust. Few comments (4 when checked).