Academics Urged to Adapt Exams and Papers for Current AI Landscape
A Brendan Nyhan post about 27 students withdrawing from an assessment turned into a sharper argument about whether old exam formats still work in the age of AI.
37987189.5K
Story Brief
In a post on X, political scientist Brendan Nyhan argued that professors should stop giving exams and papers "that are not adapted to the current AI landscape," after saying 27 students withdrew from an assessment and praising students 1, 22 and 31. The post spread through academic and tech circles, including a repost from Northwestern economist Ben Golub, and quickly became a fight over whether old exam formats still make sense when AI tools are easy to reach.
I saw this sort of thing when I taught at Caltech in about 2011. No AI back then but we had problem set scores, midterms, and finals, and that generated easily enough dimensionality to identify precisely who was in which study group and which person (there was always one) was coasting in each group.
Casey Handmer
