Every academic honor code rests on a single, fragile assumption: that a person of good faith knows when they have crossed the line. Students who slip a crib sheet into a sleeve know exactly what they are doing. So do the ones who buy a term paper or copy a roommate’s problem set. The act is effortful and deliberate.
Because it is unambiguous, the honor code can ask students to govern themselves. The line is bright. They can see it, and choose to step over it. That assumption is what artificial intelligence is now dissolving, and it is why the honor code, one of the oldest and most underappreciated institutions in higher education, is genuinely at risk.
The slippery slope of generative AI
We tend to frame the problem as temptation: AI makes cheating easy, so more students will cheat.
That is true, but it is not the greatest danger. AI makes cheating unclear. These tools do not offer a stark choice between honesty and fraud; they offer a gradient.
Ask a chatbot to suggest a topic or to sharpen a thesis. Ask it to reorganize a clumsy paragraph, then to smooth the prose and clean up the citations. At no point does an alarm sound, or a student feel the cold certainty of the crib sheet. They can tell themselves, often sincerely, that they were only getting help, the way one gets help from a librarian, a search engine or a friend.
There is a real question buried in that comparison. Students have always used tools to write: spellcheck, the thesaurus, the grammar checker, the friend who reads a draft and says that this paragraph is a mess. None of these ever threatened an honor code, because each leaves the thinking with the student.
Consider the most ordinary use of AI: a student has written their own essay, their own argument and evidence and words, and then, near midnight, pastes it into a chatbot to tighten the prose. What comes back is cleaner and no longer entirely theirs.
Did they cheat? They could not say for certain, and neither, honestly, could the code they utilized. Generative AI begins in genuinely harmless territory and keeps going, by degrees, until the thinking has quietly changed hands. The distance from the first prompt to the last is a slope, not a wall.
Can honor codes survive in the AI age?
An honor code cannot survive that. It was never built to police a gradient. It was built on the premise that a violation is apparent; a person knows that they are doing something wrong. Erase the line, and you do not get defiance, but something worse for an honor system: drift. Good students, acting in good faith, slide across a boundary they can no longer locate.
This matters because it is easy to mistake an honor code for an enforcement mechanism, a rule with a punishment attached. But an honor code is simply a community’s declaration that it trusts its members, and a wager that the trust will be repaid. Its product is people who do their own work when no one is watching, because they have come to understand that the work is theirs to do.
Across Tennessee, our academic communities have long relied on this shared trust to anchor campus culture. From the century-old traditions at Sewanee: The University of the South to the values-driven ethos at Belmont University and the student accountability framework at the University of Tennessee, Chattanooga, honor codes have served as the moral compass of higher education throughout the state.
Preserving them matters well beyond the classroom. Law school, medical school and the training of any profession are where a person practices integrity before the stakes are real, before there is a client, a patient, a court or a classroom.
The honor code is the rehearsal space. A student who learns to write their own brief when a chatbot could do it faster is building a muscle they will need later, when a senior partner leans on them to cut a corner, and no one would ever know. Hollow out the rehearsal, and the failure does not vanish. It surfaces later, at a moment when the cost is borne by someone else.
So, will honor codes survive the era of AI as a new academic year begins? They can, but not if we respond the way our instincts are pushing us. The reflex is surveillance: more proctors, more detection software, more suspicion.
That reflex is a quiet surrender. It concedes that trust is dead and replaces it with monitoring, which is precisely why an honor code exists to make unnecessary. And it is a surrender that fails even on its own terms, since the detection tools are notoriously unreliable: quick to brand an honest student, slow to catch a dishonest one.
Two ways educators can reframe the AI conversation
The harder and better path is to do the two things AI has made urgent for Tennessee university educators this fall. First, redraw the line: clearly, concretely and in plain language, so a student in good faith can once again know when they have crossed it. Ambiguity is the enemy here, and specificity is a form of mercy.
Second, and more important, instructors should make the case to students themselves; not “you might get caught,” but “this is your education, your competence, your license, your name.” That case is stronger than professors and students may realize. To trade this for a single grade with negligible effect on a career is a poor bargain, and most students, told so plainly and early enough, will not take it.
An honor code was always a bet on character. AI has not won that bet. But it has, for the first time in a long while, forced us to play the hand.
Stuart N. Brotman of Knoxville is digital media laureate and distinguished senior fellow at The Media Institute. He is the author of "Free Expression Under Fire: Defending Free Speech and Free Press Across the Political Spectrum" (2026).
This article originally appeared on Nashville Tennessean: Generative AI is blurring the line honor codes depend on | Opinion
