Exam Proctoring: We Can Do Better

Recently, use of AI has prompted loss of faith in the integrity of student essays and exams. Apparently, an ethos that “only chumps work honestly” had become pervasive. The University’s response was to institute proctored exams. This approach could easily intensify, with TSA-like screening for forbidden electronic devices, and multiple exam-cams overseeing the tests.

Alternatively, one might encourage AI in grading. It is easy to imagine AI-generated submissions receiving AI-generated evaluations. Viewing a Princeton education as a commodity, bilateral application of AI could be a productivity breakthrough: Completing a semester’s submissions and exams could be reduced to a few weeks. Start-to-finish graduation requirements for the 4-year program could be done in a year, with many fewer faculty. There would be opportunity for huge tuition reductions.

With either approach, however, something important is lost. Neither enhanced policing nor enhanced AI use improves the quality of the academic enterprise. After graduation, many will go on to professional careers (e.g. medicine, law, science, journalism) in which working honorably is expected by patients, clients, and colleagues. If the undergraduate experience disparages integrity as quaint or unrealistic, then when does honesty become the performance standard?

I would welcome an effort to rehabilitate Princeton’s culture of honor. A century ago, adoption of lecture course precepts showed the University valuing student-faculty interaction. Perhaps configuring courses so that students spoke regularly with faculty, who knew their names, might strengthen the expectation of honesty. Certainly, Princeton should be able to devise a more positive remedy than enhanced policing.