The day before a take-home exam at Wharton, a few students warned their professor that classmates were planning to cheat with AI. The professor was Adam Grant, probably the best-known organizational psychologist in the country, and what he did next is the interesting part. He scrapped the exam and replaced it with a single essay question: use the organizational psychology you've studied to explain how to motivate students to stop cheating.
Ninety-five students answered, and their responses dismayed him. Hardly any treated cheating as an integrity violation. The majority made excuses: everyone does it, the stress is enormous, there's no choice. Grant compares the mindset to baseball's steroid era, and his survey research from elite campuses goes further. Students now admit to cheating with their full names attached, and one Ivy League student even told him: "Cheating is a creative expression and a form of problem solving."
Grant's conclusion is a direct challenge to advice we published a few weeks ago. In-class essays and oral exams, he argues, are just obstacles, and obstacles are never enough; the durable fix is building students' moral identity, their sense that integrity is part of who they are. So, was our assessment-redesign issue wrong? Nope, and I don't think he's necessarily wrong either. The blend is where this gets useful for your district.
Some folks have reached out to us asking how to subscribe themselves or their colleagues to this newsletter. Click here to do so. We send a newsletter every two weeks about the latest happenings with K-12 and AI.
IN THIS ISSUE:
The evidence that ethics can actually be taught, and what kind of teaching works
Why the AI-open assignment may be the best moral identity practice a school can offer
95
Students in Grant's Wharton course whose exam became an essay on why students cheat. Source
1.2 million
Investment advisers studied in the landmark research on whether ethics training changes behavior. Source
25%
How much less likely advisers who trained on ethics-heavy exam content were to commit misconduct, compared to peers at the same firm. Source
10 of 21
Honesty-oath variations that measurably reduced dishonesty in a Nature Human Behaviour megastudy. The wording did the work, not the signature. Source
1 in 3
Students who have already cheated by elementary school, per the anonymous survey research Grant cites. High school ethics units arrive years late. Source
The strongest study in Grant's argument deserves your attention, because it answers the question every skeptic asks about character education: does it change behavior, or just talk? Researchers at Notre Dame, MIT, and the London School of Economics studied 1.2 million investment advisers in the wake of a 2010 change to the licensing exam that shifted the exam's weight between ethics and technical content. Among advisers at the same firm, in the same office, in the same year, those who trained on the ethics-heavy version went on to commit misconduct at a rate that was one-fourth lower. The mechanism wasn't rule memorization: the researchers found the training changed advisers' perception of what conduct was acceptable. And when their employers got caught in scandals, the ethics-trained advisers were more likely to leave.
The other useful evidence is about how not to do it. Honor pledges, as most schools run them, are a signature ritual, and the research on them is mixed at best. A Nature Human Behaviour megastudy tested 21 versions of an honesty oath and found only 10 reduced dishonesty at all. What separated the versions that worked was content: oaths that made the moral stakes explicit, why honesty matters, and who dishonesty harms, performed best, with the strongest nearly halving cheating. The signature line at the bottom of your student handbook is probably doing nothing. The paragraph above it, if anyone wrote it with care, might be doing a lot.
OUR TWO CENTS
Here's where I part ways with Grant, a bit. He files in-class essays and oral defenses under "obstacles," and if that's all a district builds, he's right. A wall teaches a kid nothing except where the wall is.
But there's one redesign that isn't a wall, and it's the one we've argued for all along: the assignment where AI is required, openly, on the record. Consider what that assignment actually asks of a student. Disclose what you used. Show what the AI contributed and what you changed. Defend the judgment calls as yours. (Effectively… show your work.) That's not an obstacle to cheating; there's nothing left to conceal. It's a rep of exactly the muscle Grant wants built. The advisers in that study didn't internalize ethics from a lecture; they internalized it by grappling with ethical material until it changed what they saw as acceptable. An AI-open assignment makes a student grapple with the same temptation every week, the one their generation faces.
We do a small version of this ourselves: the line at the bottom of every issue disclosing that we use AI to polish our content. It costs us nothing, and it isn't nothing. Integrity, practiced in small, visible ways, becomes identity, and identity is what shows up when nobody's looking.
Keep the redesign work from our integrity issue, and add the motive layer described in this issue. They're not rivals.
Make disclosure the default on AI-open assignments: what I used, what it gave me, what I changed, what's mine.
Grade the judgment, not just the product, so that the disclosed process earns points and honesty pays off.
Have teachers model it. A teacher who discloses the use of AI in building a lesson is teaching integrity without a lecture.
If the signature ritual does little and content does the work, most districts are one rewrite away from a more useful honor code. The megastudy's lesson translates directly: an integrity statement should say, in plain language, why honesty matters here, that cheating is unfair to the students who didn't cheat, and that trust, once lost, is expensive to rebuild. Students should encounter it where the temptation lives, on the assignment itself, not once in August inside a handbook acknowledgment nobody reads.
Timing is the other half. Grant cites anonymous survey research finding that about a third of students have cheated by elementary school, with rates climbing through middle and high school. The adviser research points in the same direction: ethics training worked best on the inexperienced, the people whose sense of acceptable conduct was still forming. A district whose integrity push starts in ninth grade is doing remediation. The one that starts in third grade, with age-sized versions of the same question: what did you make, what helped you, what's yours, is doing formation. And this connects to something your students already told the country: the 98 students who wrote the national AI framework this summer put guardrails in place for themselves first. Kids will hold a standard they helped set. That instinct is moral identity, and it's sitting in your buildings waiting to be asked.
OUR TWO CENTS
Three moves, none of which require a new program. First, rewrite your integrity pledge for content, and consider retiring the signature theater: a short paragraph on the assignment that names why honest work matters beats a legal acknowledgment in a handbook. Second, push the conversation down the grades. Elementary students can absolutely discuss what's fair when a tool helps them, and, according to research, that's the age when such discussion shapes who they become. Third, close the loop adults control: when students see teachers disclose AI use, see appeals taken seriously, and see effort graded honestly, they conclude that integrity is how this place works. Culture is just the pattern of what people see. Make the pattern honest and visible, and the pledge starts describing something true.
Rewrite the honor pledge for moral content: why honesty matters, who cheating harms. Cut the signature ritual if it's all you have.
Put a two-line integrity statement on assignments themselves, where the temptation actually lives.
Add age-sized integrity discussions to elementary AI literacy, before habits form.
Audit what students see: teacher AI disclosure, working appeals, honest grading. The visible pattern is the curriculum.
Think Again: The Power of Knowing What You Don't Know by Adam Grant
Grant wrote this issue's anchor essay, and his best-known book is about the skill underneath it: letting go of conclusions that no longer fit the evidence. The final section is aimed at schools and at building cultures of rethinking, which is a fair description of what the AI era demands of every district office. We'd argue this issue is an exercise in it: Grant's essay made us rethink a piece of our own advice, and the result was better than either position alone. Useful for any leadership team considering a rethink of an honor code.
Chat with ClassCloud
We’re listening. Let’s Talk!
This newsletter works best when it’s a conversation, not a broadcast. If you want to talk through how any of this applies to your district specifically—or if you have feedback on what would make this more helpful—just hit reply. We read and respond to everything.
Schedule a Virtual Meeting
Thanks for reading,
Russ Davis, Founder & CEO, ClassCloud ([email protected])
Tracy Walters, Director of Customer Growth ([email protected])
ClassCloud is an AI company, so naturally, we use AI to polish our content.
