The Biggest Questions About AI
The map · 5 Humanity · 5.2 Education and human cognition · 5.2.3

Assessment

How can schools and employers determine what a person genuinely knows or can do?

Take-home assessment is broken — Mollick's 'Homework Apocalypse' arrived on schedule. The redesign question is what to verify (process? performance under observation?) and at what cost to learning itself.

View on the map → · Open in Browse →

What changed
March–August 2026 · swept August 3, 2026 · editorial review pending
01

The credential-validity problem got measured and the responses began: a 20-university, 95,000-student study found about a third regularly use AI on assignments and 9% admit outright cheating, and the visible reaction has been the return of blue books and oral exams — each verifying authorship at a real cost to what it can measure and to equity.

Recent thinking
3 pieces · March–August 2026 · all 3 chronologically →
Patricia Waldron · Cornell Chronicle · 21 May 2026 news
Widespread AI misuse means higher ed must rethink assessment
The fact that students are misusing GenAI is a problem for assessment validity, and that's a problem for the credibility of university credentials.

Reports the Kizilcec, Chirikov & Smirnov study (Science, 2026) across 20 universities and 95,000+ students: about a third regularly use GenAI on assignments and 9% admit cheating — the largest measurement grounding the 'what does the credential certify' problem.

Josephine Walker · Axios · 14 Mar 2026 news
Blue books make an "out of step" campus comeback in the AI era
Colleges besieged by AI-generated writing brought back blue-book exams to deter cheating, but some educators say hand-written tests don't showcase students' best work and disadvantage swaths of learners.

Documents the return of blue books as the live assessment response and surfaces the tradeoff: in-class handwriting verifies authorship but may under-measure ability and disadvantage some learners.

Hengzhi Hu & Harwati Hashim · Educational Researcher · Mar 2026 paper
Oral Assessment in the AI Era—Equity, Validity, and Scale
Operationalize equity by separating language production from disciplinary reasoning and providing targeted accommodations

Peer-reviewed treatment of the other main verification response — oral exams — taking the equity, validity, and scalability objections seriously rather than treating orals as an obvious fix.

Foundational reading (2)Assessment reform for the age of artificial intelligenceLodge, Dawson et al., TEQSA · 2023The Homework ApocalypseEthan Mollick · 2023
Previous5.2.2 Learning versus outsourcingNext5.2.4 Cognitive inequality