For students wrongly accused of using AI
Falsely Accused of Using AI? You're Not Alone, and You Have a Defense.
Being flagged by an AI detector is more common than universities admit. If you did not use AI, your writing process is the evidence. The Defense Assistant walks you through what to gather, what to write, and what to say — calmly and in about ninety seconds.
The short answer
If you didn't use AI, your strongest evidence is your writing process. Independent studies put AI-detector false-positive rates at 5 to 20 percent. Stanford researchers found 61.3 percent on essays by non-native English writers. Vanderbilt University disabled Turnitin's AI detector in August 2023, estimating it would have wrongly flagged 750 papers per 75,000 submissions.
Where to start
The eight chapters below cover everything: the first hours after the accusation, the evidence that has cleared students in named cases, the verified false-positive data on every major detector, the appeal-letter structure, and your due-process rights.
- 01 Read →
Falsely Accused of Using ChatGPT
The seven-step playbook for the first 72 hours. What to do, what to gather, what to say.
- 02 Read →
How to Prove You Didn't Use AI
The five strongest forms of evidence, with step-by-step instructions for Google Docs, Microsoft Word, Notion, and pen-and-paper writers.
- 03 Read →
Are AI Detectors Accurate?
Independent studies, vendor admissions, university decisions, and the demonstrations where AI detectors flagged the "I Have a Dream" speech.
- 04 Read →
How to Appeal an AI Cheating Accusation
Your due-process rights and the four-part structure of a strong appeal letter, with examples and the resources to contact.
- 05 Read →
Turnitin Said I Used AI, But I Didn't
Turnitin's own admissions, Vanderbilt's decision to disable the tool, and what to do if Turnitin is the only basis of your accusation.
- 06 Read →
AI Detector False-Positive Rates
Every published false-positive rate, with sources. Turnitin, Stanford, independent studies, and the scale arithmetic that shows the absolute harm.
- 07 Read →
Professor Accused Me of Using ChatGPT
The first-24-hours playbook. What to send, what NOT to send, and how to preserve the evidence before it changes.
- 08 Read →
Your Rights When Accused of Using AI
Due-process protections, the federal lawsuits where students fought back, and the free legal-help organizations to call.
Frequently asked
Are AI detectors accurate?
Not reliably.
Independent academic studies report false-positive rates of 5 to 20 percent. Stanford researchers (Zou et al., 2023) found a 61.3 percent false-positive rate on essays by non-native English writers. Turnitin's own documentation admits a 4 percent sentence-level false-positive rate. Read the full breakdown.
What is the strongest evidence I can provide?
Your writing process.
Time-stamped Google Docs version history, Microsoft Word AutoSave timelines, or playback extensions like Draftback. Two named UC Davis cases were resolved when students showed edit histories spanning hours. Step-by-step guide.
Should I confess to lower the punishment?
No. Not if you did not use AI.
An admission carries downstream consequences far more serious than the original accusation, including referral to a permanent record, visa issues for international students, and complications for graduate-school applications. Gather your evidence first.
Has any university disabled Turnitin's AI detector?
Yes — Vanderbilt, August 16, 2023.
Vanderbilt University disabled Turnitin's AI detection tool citing lack of transparency, bias against non-native English speakers, and an estimated 750 false positives per 75,000 papers. What this means for your case.
Can I sue if I am wrongly accused?
Some students have, and some have won.
At least five federal lawsuits have been filed since 2024. Orion Newby won a landmark ruling against Adelphi University in February 2026, with the judge calling the AI plagiarism finding 'without merit.' Consult an attorney before filing. Read about student lawsuits and rights.
What if English is not my first language?
Cite Stanford's 61.3 percent finding in your appeal.
The Zou et al. (2023) Stanford study found that AI detectors falsely flagged TOEFL essays by non-native English writers 61.3 percent of the time on average, compared with near-zero false positives for native US 8th-grade essays. This is peer-reviewed evidence of bias.
The standard we hold ourselves to
Every number, every named case, and every percentage on this site carries its source. We checked Turnitin's documentation, Stanford's 2023 study, Vanderbilt's August 2023 announcement, and the public court filings for every lawsuit we cite. The full list is on the sources page. The site is non-affiliated with any AI-detector vendor. We do not run ads or affiliate links.
Everything here is informational and is not legal advice. If your visa, scholarship, or expulsion is at stake, contact a licensed attorney or your campus ombuds. The Defense Assistant lists free resources by jurisdiction.