On the same week that Trump dismissed the entire National Science Board, Nature published a deep dive into how decision-making became data-driven, charting the long arc of the evidence revolution. The timing is almost too cruel. The body that exists to advise a president on scientific evidence is gone. The evidence revolution, it turns out, has no backup copy in government.

The Parallel Audit in the Art World

Meanwhile, galleries and auction houses are integrating AI into provenance research, attribution, and market analytics. The art world has always had a shaky relationship with evidence. Connoisseurship, that polite word for vibes-based attribution, is finally being pressured by machine pattern recognition. The AIPAD photography fair, per Hyperallergic's review, pushed back against AI image generation this year by doubling down on craft and process. In a room full of Latin American photographers insisting on materiality, the epistemological stakes were clear: authenticity requires a chain of custody that algorithms can disrupt but cannot replace.

Evidence, Expertise, and Who Gets to Say So

The NSB dismissal and the AI-in-galleries story share an uncomfortable structure. Both involve the delegitimization of human expert bodies in favor of faster, cheaper, more legible systems. The NSB had 25 members, each an expert. An AI provenance tool has millions of training examples. Scale is being mistaken for authority. A 2022 paper in Science by Lazer et al. warned that data-driven governance can produce high-confidence wrong answers when training sets encode historical bias. The evidence revolution, it turns out, can be used to dismantle evidence just as easily as it builds it. . The race to control what counts as true is also a market.