Grades dropped from 96 to 48 percent when a Brown professor made students take the exam without AI
2026-07-13
Summary
A Brown University professor noticed unusually high scores on a take-home exam, suspecting students used AI tools like ChatGPT to cheat. When he conducted a proctored, in-person final exam, average scores plummeted from 96% to 48%, confirming his suspicions. This issue reflects a broader trend observed in two studies that show students performing significantly worse on supervised exams compared to unsupervised assignments, likely due to reliance on AI for homework.
Why This Matters
The situation at Brown University highlights a growing concern about academic integrity in the age of AI, raising questions about how technology might be impacting education quality. As AI becomes more accessible, educational institutions may need to adapt their assessment strategies to ensure fair evaluation and maintain academic standards. This issue also has broader implications for workforce readiness, as reliance on AI could undermine essential problem-solving skills.
How You Can Use This Info
For educators, this highlights the importance of developing assessment methods that accurately reflect student capabilities, potentially incorporating more in-person evaluations. Employers should be aware that academic achievements might not always fully represent a candidate’s abilities due to the potential misuse of AI. Professionals can also consider the ethical implications of AI in their fields, promoting a culture of integrity and transparency.