Generative AI availability during coursework compresses university grade distributions, eroding grades' value as a hiring signal

Hausman, Rigbi, and Weisburd tracked roughly 36,000 students across roughly 6,000 courses at a large Israeli university from 2018–2024, comparing AI-compatible courses (take-home work) to AI-incompatible courses (in-person exams) before and after ChatGPT's introduction. Their headline finding: "AI availability raises grades, especially for lower-performing students, and compresses the grade distribution, eroding the signal value of grades for employers."

Quantitatively, grades in AI-compatible courses rose "about 0.6–1 points" on average against a mean grade of 86, with 25th-percentile students gaining "more than 2.5 additional points" in the most controlled specification. ChatGPT's introduction was "associated with a 33% drop in the share of students failing AI-compatible courses in the first year, and a 50% reduction in the share receiving a D grade in AI-compatible courses the second year." The authors frame this as a signaling problem directly: grades increasingly "[pool] students of a variety of abilities under similar grades," which "complicat[es] hiring, forcing employers to find other means of assessment" — the mechanism by which employers respond to a degraded signal by seeking new, costlier ones.

This is the clearest evidence in the vault that the cost-of-production collapse extends to formal academic credentials specifically, distinct from freelance-market reputation (claim-chatgpt-release-cut-upwork-writing-freelancers-jobs-and-pay) or hiring signals generally (claim-llm-collapse-of-costly-writing-signal-cuts-meritocratic-hiring). It parallels claim-szabo-collectible-monies-collapse-when-costliness-becomes-forgeable: a grade, like a wampum bead, held value because it was costly to fake; once that cost fell, so did the value of the signal itself, regardless of the underlying ability it was meant to certify.