In an email to undergraduates on Wednesday, Harvard College Dean David J. Deming shared his priorities for the coming year, including both the cultivation of a “learning community” and clarification of artificial intelligence policies across the school.
On these issues, the College’s highest office apparently hopes to cheapen a Harvard education out of fear that some students will be careless with their own learning. But principles should not be adjusted to accommodate those who violate them.
Deming wrote that AI bans are “unenforceable” and contribute to creating a temptation for students to cheat so as to avoid “falling behind.” In some cases, his suggestions address the AI-cheating problem by removing AI from the picture through in-person written or oral exams. On these points, I agree with him — in-classroom accountability is a necessity under the AI threat.
However, Deming seems to bend the knee to the perverse incentives AI creates. Due to both the temptation students may feel to use the technology and the distrust it fosters between faculty and students, he proposes that Harvard should strive for “AI acceptance, or even encouragement” in writing-heavy courses.
This stance is strange given the manner in which the College handled the grading issue last year. Student complaints about the cap on A’s often centered around perverse incentives, namely decaying collaboration and further gamification of scheduling. But the College, in spite of these threats of student behavior, remained steadfast in supporting what they perceived as the best path for the long-term intellectual health of the University.
Deming himself even said at an open forum in March that he believes students will set aside the cutthroat pressures and continue to collaborate. When it comes to grading, Deming seems happy to legislate against the grain of the cynical, optimizing student mind.
I wrote last year about the fundamental incompatibility of artificial intelligence with a humanities education. Many prominent faculty members have shared similar thoughts in response to Deming’s email. English professor Deirdre Lynch highlights the importance of the dialectic between thought and language in writing, and History professor Peter E. Gordon rightly worries AI usage may limit thought to the established pools of knowledge on which AI models are based, thus limiting autonomous thought.
I will not rehash these arguments further, but they are among the reasons that I believe the incorporation of AI use in humanistic writing to be a death knell for an intellectually formative and worthwhile education. What’s worse is that this death would be caused by those who, by outsourcing the writing process to AI, showed limited regard for their education to begin with.
In his email, Deming links to a Substack post from Colorado State University professor Kyle L. Saunders in which he explains that, when he assigns out-of-class work, AI use is encouraged at any stage in the writing process. Saunders, like Deming, is partly motivated by a fear of “punishing” students who choose to do their work honestly. The posited solution to this perceived injustice is then to encourage all students to substitute their own words for words given to them.
If Deming truly hopes to build the “learning community” he heralds in his email, the absolute last thing he should do is legitimize cutting intellectual corners by advocating changes that accommodate such behavior.
If AI prohibition is indeed unenforceable, then a number of students will cheat and get away with it. Let them! The loss is their own. Free thought, growth through struggle, and pride in work are not ideals we should so freely sacrifice on the altar of ease and efficiency.
The fact that so many Harvard students are willing to do so is a problem that should be confronted directly (in admissions, perhaps?) rather than begrudgingly endorsed. It is discouraging that the College’s top official should seem to take such a pessimistic view of how much integrity can be expected from Harvard students.
I suspect I might not be alone in feeling that after a summer away from Cambridge, the amount of energy and animosity spent on whether an A or an A- would be the standard was, to an extent, much ado about nothing. The effect of such a change on our lives has likely been overstated. When it comes to AI, though, the existence and quality of student thought could not be more central in the power of a Harvard education. But on this issue, in the face of bad behavior, Deming has thrown his hands in the air.
Incentives are not mandates — it would do Harvard students, faculty, and administrators alike a lot of good to internalize this fact. In general, but especially with AI, discussions of “the future” treat tomorrow as an inevitability foisted upon us rather than a contingency actualized by good or bad decisions. Admittance to Harvard College should come with the rigid expectation that one has the resolve to do what is right in the face of what is easy.
It would serve us well if our Dean thought the same.
Adam N. Chiocco ’27, an Associate Editorial Editor, is a Philosophy concentrator in Pforzheimer House.
Want to keep up with breaking news?Subscribe to our email newsletter
Have a tip for The Crimson?Share it confidentially
