Professor denounces mass AI fraud on an exam at Brown
Widespread use of AI tools to complete a “closed-book” take-home exam at Brown University is prompting broader questions about academic integrity and whether traditional assessment models still work. Commenters argue that take-home exams and curve grading now virtually incentivize cheating, especially when degrees function mainly as job-market credentials and universities are reluctant to enforce penalties. Many see a shift toward in‑person, proctored, often handwritten or oral exams as inevitable, while others advocate redesigning courses so AI use is either irrelevant or explicitly integrated into how students are taught and evaluated.
Exam design & AI-enabled cheating
- Many argue the real problem is giving “take‑home, closed‑book” exams at all; such formats assume an honor culture that no longer exists.
- AI is seen as lowering the barrier and increasing the scale of cheating: fast, accessible, hard to detect, and capable of doing entire assignments.
- Others note cheating on take‑home work was already widespread pre‑AI; AI changes magnitude and convenience, not the basic behavior.
- AI‑detection tools are widely viewed as unreliable and dangerous (false positives, opaque methods).
Student incentives and culture
- For many, college is a credential and networking tool, not primarily about learning. The degree is a “meal ticket,” especially at elite schools.
- Grade curves and competitive programs create prisoner’s‑dilemma pressures: if classmates cheat, honest students risk lower relative ranks and worse career options.
- Post‑COVID and with rising costs, several commenters see cheating as normalized and stigma eroded; students “reward‑hack” any metric.
Ethics, integrity, and curves
- One side insists cheating is always a choice; widespread dishonesty in society doesn’t justify it.
- Others argue it’s naive to demand high integrity from students while institutions and elites behave cynically.
- Curved grading is heavily criticized for turning learning into zero‑sum contests that structurally incentivize cheating.
Institutional responses and proposals
- Strong push for in‑person, proctored exams: on paper or on locked‑down institutional computers; some support handwritten work, others say typing with controlled devices is enough.
- Suggested tools: testing centers with sandboxed PCs, oral exams, one‑on‑one interviews about submitted work, frequent low‑stakes quizzes, project‑based and in‑class work.
- Some instructors already design courses “adversarially,” ensuring that the easiest path to a high grade still requires genuine understanding.
- Administrations are described as reluctant to confront cheating because of enrollment, funding, and reputational incentives.
Debate over grading and purpose of university
- Several question whether grades meaningfully predict job performance; many report employers rarely care beyond “has a degree.”
- Some professors doubt the value of grading at all, seeing it as unpaid screening for employers amid pervasive grade inflation.
- Others defend grades as necessary feedback, prerequisites for advanced courses, and a core part of a credential that must retain meaning.
International and historical perspectives
- Commenters contrast systems: weed‑out pen‑and‑paper finals in parts of Europe, oral exams in Hungary/Italy, honor‑code take‑homes at some US schools.
- There is skepticism that older honor‑code models can survive in the current environment without substantial cultural and structural change.
Views on AI’s role in education
- One camp wants strict prohibition in assessments and sees AI cheating as hollowing out expertise and devaluing degrees.
- Another argues AI use is inevitable in work; education should teach students to use AI well, assess understanding via interviews, presentations, or process‑focused tasks, and separate “no‑AI” fundamentals from “AI‑assisted” real‑world skills.