Chronicle: Professors Retire Early, Citing AI-Driven Cheating

The most vivid piece of this week's higher-ed AI reporting is a Chronicle of Higher Education profile by Adrienne Lu on professors who have decided the essay, as they knew it, is not worth defending. Michael Hoberman, a Fitchburg State English professor of 25 years, is retiring in December at 62, a few years ahead of his plan, after trying and failing to find a workable answer to students slipping AI-generated sentences into their writing. His workaround was to move most writing into class with pen and paper, which did not give students enough time or him enough material to comment on for them to grow as writers. He says he is leaving the fight to "younger, more imaginative, resourceful people" because "I don't have it in me to fight that battle."

He is not the only one in the piece. Allison Cummings, an English professor at Southern New Hampshire University, retired in the spring at 61. Kari Kalve at Earlham College moved her retirement from 67 to 64. Jessica Wolfe, 56, left tenure and a chaired professorship at UNC-Chapel Hill. Paul Robichaud, 54, at Albertus Magnus, is uncertain he wants another decade of it. Lu anchors the trend with a Center for Retirement Research figure: after ChatGPT launched, 16 percent of postsecondary educators aged 55 and older left employment within a year, up from 11 percent before.

For our readers this slots into a run of education coverage where the assessment problem keeps outrunning the fix, from Denmark adding oral defenses to upper-secondary exams to individual faculty deciding the redesign is a multi-year project they will not live to see through. The essay is the load-bearing tool of humanities teaching, and when detection is unreliable, retire is a rational answer for anyone already near the door. Five experts in our Who's Who directory shared Lu's piece this week.

A few things Lu's piece does not settle. The 16 percent number is a one-year post-ChatGPT window, not a causal claim, and a lot else was moving in higher ed at the same time. Five named professors is a strong opener but not a base rate, and the article does not say how many retirements are AI-driven versus AI-labeled after the fact. It also does not say who backfills these tenure lines, which is the number that decides whether this becomes a staffing crisis or a generational handoff.

If there is a hopeful reading, it is Hoberman's own: some of this work needs "younger, more imaginative, resourceful people" willing to rebuild assessment from scratch. Who pays for that redesign, and how fast, is the next question.