TeacherAware: Provenance for the Classroom —
Re-Reversing the Cheating Incentive

Aaron Kushner
2026 · Education · Provenance · K-12 Integrity
Developed with: Claude (Anthropic)

"I watched a freshman I knew sign the declaration that he'd done his homework without A.I. as ChatGPT was still open in the next window" — Theo Baker, Stanford senior, The New York Times, May 17, 2026. In April 2026, Stanford reinstated in-person proctored exams after a century-long ban — abandoning an honor system that AI had made structurally untenable.

The Problem: The Institutional Collapse

The point was never to make cheating impossible. It has never been impossible — before AI it was merely difficult and risky: a ghostwriter cost money, a copied paper could be recognized, and the effort of cheating well approached the effort of doing the work. What AI changed is the incentive structure. Cheating became easier than not cheating — faster, better, available at 2am, and undetectable by any existing system — while the reward (grade, credential, competitive advantage) stayed the same. For a century, Stanford University maintained an honor system premised on trusting students not to cheat. In April 2026 it reinstated in-person proctored exams for the first time in that century. The honor system did not collapse because students became less honorable. It collapsed because the cost-benefit of cheating inverted overnight, and rational students responded rationally to an irrational incentive structure.

The institutional responses on offer are both losing propositions. Post-hoc AI detection is structurally unwinnable — probabilistic, biased against certain writing styles and native languages, and locked in an arms race with humanizers and evasion tools that improve as fast as the detectors do; even detection vendors concede the race. In-person proctoring works but is expensive, unscalable, and returns the classroom to the nineteenth century — it does nothing for the take-home essay, the research paper, the weeks-long composition process where writing is actually taught. Faced with that choice, teachers have quietly stopped assigning take-home writing at all. That is the real cost: not the cheating that happens, but the assignments that no longer do.

What every response so far shares is an absence: there is no record of the writing process. Detection operates on the finished page and guesses. A student who drafted 1,200 words over three evenings, revised them a dozen times, and used a sanctioned AI tutor for two clarifying questions produces a document indistinguishable — to any detector — from one generated in thirty seconds and retyped. The record that would distinguish them does not exist. TeacherAware creates it.

We Present TeacherAware, the Solution

TeacherAware is a writing environment for the classroom that records process, not suspicion: keystrokes, timing, revisions, paste events, and sanctioned AI use are sealed into a SHA-256 hash chain as the student works, checkpoint by checkpoint, and anchored to GitHub at submission — a timestamp neither student nor school can rewrite. The result re-reverses the incentive: honest work becomes the cheap path again (write in the workspace, and the record makes your case for you), while cheating becomes expensive (defeating a keystroke chain inside a proctored session is more work than writing the essay). Every output is advisory — evidence for a human teacher, never a verdict about a student.

The same environment that records unauthorized AI use can host authorized AI — a constitutionally-constrained writing collaborator, used in the open, on the record. Integrity infrastructure and AI-literacy infrastructure turn out to be the same infrastructure. TeacherAware is free, requires no school account and no per-student cost, and its credential is self-verifying: the math either resolves or it doesn't, with no trust placed in us.

The Architecture A Self-Verifying Credential, Cosigned by GitHub

The whole product rests on one move: a SHA-256 hash chain of every writing session, anchored to GitHub at submission time. The chain is self-verifying — the math either resolves or it doesn't — and the timestamp is borrowed from infrastructure neither student nor school can rewrite. No server of ours. No trust placed in us. The credential travels with the assignment.

What a submission produces — self-verifying, GitHub-cosigned
Assignment write TEACHER AWARE Any browser Recording Workspace { chain: a3f9…c12e 7d2a…88bf } credential.json submit GitHub Timestamped commit Unalterable record Cloud record TEACHER / SCHOOL ✓ verify the work how · when · with what help STUDENT ◎ rebut false accusations tamper-evident process log
The credential travels with the assignment — attached to the submission, readable by any teacher in any browser. It delivers two things at once: rapid teacher verification — a clear record of how the work was written, when, and with what sanctioned help, readable at the dashboard in seconds; and a foolproof defense for the honest student — a tamper-evident process log that cannot be constructed after the fact, giving every student standing against a false AI accusation.
↓ try it
The commercial front door — live. This is the actual application running in this page. "Know the work is theirs": the landing page a teacher or district hits first — value proposition, how it works, privacy posture, and the tier story, ending at the demo.

References

  1. Baker, T. (2026, May 17). What A.I. Did to My College Class. The New York Times.
  2. Goldstein, D. (2026, June 16). How Students Are Cheating With A.I. — and the Tools That Promise to Catch Them. The New York Times. (Tool-by-tool coverage and the layered-defense answer are walked in the companion appendix, AI Cheating at a Glance.)

The Document Package

This overview is the front page. The depth lives in the companions: the position paper (why detection loses and provenance wins), the go-to-market brief (the free foundation wedge), the as-built spec sheet (every shipped feature), and the cheating-tools appendix (the adversary roster, tool by tool). The person behind it: Aaron Kushner — profile & résumé.

Author Contribution Statement

The author used AI extensively in this project and this communication.