Proof of Effort
Proof of Effort
A Forensic Tool that Keeps the Conversation Honest
Proof of Effort for Educators: A New Standard in Academic Integrity
A Forensic Tool that Keeps the Conversation Honest
The question facing your institution has changed. It's no longer "Is this AI?" — a question no tool can answer reliably — but "Was a human genuinely engaged in producing this work?" Proof of Effort answers that one directly. It gives students a way to show their effort and gives faculty independent, verifiable evidence to rely on — without surveillance, and without the false accusations that current tools produce.
Why Not Just Use an AI Detector?
Because detectors put your institution at risk, not just your students. They flag a meaningful share of genuinely human work as AI-generated, and that risk falls hardest on non-native English speakers, who are disproportionately misflagged. When a student disputes the result, you have nothing independent to compare to. And the approach is losing ground: as AI writing improves, AIdetection accuracy falls.
What does PoE Give That Detectors Don't?
Defensible evidence instead of a contestable verdict. A detector hands you a probability you can't explain and can't defend when challenged. PoE hands you a sealed, examinable record of how the work was produced — the pauses, revisions, and reconsideration of genuine composition that faculty can verify independently and that a student can stand behind. It does this efficiently on one page of analysis that does one thing, prove the student's effort.
How Does It Work for a Student?
PoE records how the work was done, not what the student wrote. The student opens their assignment in LibreOffice Writer or Microsoft Word with the PoE extension installed and writes as they normally would — no extra steps. PoE observes the rhythm of the session; the data is sanitized as it's captured, so nothing identifying is ever kept; and the student submits the resulting proof inside their document. Think of it like a fitness tracker: it records how you moved, not what you said or where you went.
Does this Surveil Students or Put their Privacy at Risk?
No. PoE is built to protect students, not monitor them. It never reads or stores the words a student types, never records content, and never runs in the background between writing sessions. It captures no biometric data and cannot identify anyone — raw timing is destroyed, and only non-identifying aggregates survive. The evidence belongs to the student, works identically regardless of writing style or language proficiency, and is designed around FERPA and GDPR data-minimization principles. There is no "fingerprint" and the data is tamper evident.
Can't a Student just Retype AI Text to Defeat It?
Composing and transcribing (retyping) leave different behavioral patterns. Original writing is recursive and full of correction; retyping is more linear. PoE examines a document section by section and flags those sections that measure as transcribed. The quality of the retyped text doesn't change the signal, and reproducing authentic-like effort, convincingly, takes more work than writing original text. Also, simply deleting text doesn’t delete the record. The trial and error needed to train oneself to fake an authentic signal is prohibitive and would need to be done on fresh documents 600 words at a time.
Won't PoE Falsely Accuse Students, the way Detectors do?
No — because PoE makes no accusation. It produces a measurement, not a verdict: a record of the effort in a session, on a spectrum between Authentic and Transcribed. PoE shows what the writing process looked like and leaves the judgment to you. The evidence informs your academic-integrity process; it does not substitute for it.
What Makes Transcription Flags Different from an AI Detector’s Verdict?
AI detectors give you a probability against an entire document or sentence by sentence. The former is too broad especially, when the result is a guess based on words on a page. The latter is just as much a guess. PoE resolves the tension between too much and too little cleanly. Transcription flags aren’t a guess based on finished work, but a measurement of the effort captured in action. Low effort work leans closer to Transcription, and High effort work closer to Authentic original compositon. By dividing up the document into sections of at least 600 words, PoE ensures it has enough data to take a reliable measurement.
Do Transcription Flags Mean the Student Cheated?
Not necessarily. It’s just a data point. A highly reliable data point, but one that must be taken in context. This is why each transcription flag comes with a confidence level (Low, Med, High). PoE is a writer focused system that gives the benefit of the doubt to the writer, whenever possible. Because transcription signals are so distinct across many dimensions, the program is able to provide a confidence level for each flag. Many flags at high confidence says something different than one low confidence flag among many non-flagged sections. This allows the system to account for human variation without becoming a blunt and binary verdict machine.
What Does it Mean if Some Sections are Flagged and Others aren’t?
Authentic writers often quote long passages from other texts or can enter extreme flow states where their writing is so clean and error free that it’s almost like they’re taking dictation. Because a transcription flag can only be a product of sustained linear and unusually perfect input, the likelihood of this occurring across multiple writing sessions, and multiple sections is very low. PoE provides you with this data at a glance on one simple scoring page that the student also saw before submission. The conversation you’ll be having with them will be based on mutually understood data and not surprise verdicts built on guesses.
Is It Hard to Deploy?
No. PoE installs as a lightweight extension to word processors students likely already use, with no IT infrastructure changes. Students submit the proof through your existing systems, the same way they submit any document, and can check their own score before submitting. Faculty verify on their own computer in seconds using the same free extension or through our website. Both parties see the same result, which is what removes the doubt.
Is the Method Academically Sound?
Yes. PoE rests on decades of peer-reviewed writing-process research establishing that authentic composition produces measurable behavioral patterns that reflect genuine cognitive engagement. The PoE patent pending method has been tested on the same samples the research was built on but produces data that can't be used to identify the writer or reproduce their content.
Research references are available on request to establish the scientific basis for your integrity committee.