Sources.
Every numerical claim, named case, and percentage on this site carries a primary source. Below is the full list, with the URL and the anchor where each source is cited. If you find a stale link or an updated primary source, please write to the contact address on the disclaimer page.
Universities and institutional decisions
- Vanderbilt University — Brightspace Support: "Guidance on AI Detection and Why We're Disabling Turnitin's AI Detector" (August 16, 2023)
- vanderbilt.edu/brightspace/2023/08/16/guidance-on-ai-detection
- Cited as: A1 — Vanderbilt 750/75K. Used in pillars: 1, 3, 5, 6.
Peer-reviewed academic research
- Stanford HAI — "AI-Detectors Biased Against Non-Native English Writers" (Zou et al., 2023)
- hai.stanford.edu/news/ai-detectors-biased-against-non-native-english-writers
- Cited as: A2 — Stanford 61.3% TOEFL false-positive rate. Used in pillars: 1, 3, 4, 6.
- The Markup — "AI Detection Tools Falsely Accuse International Students of Cheating" (cross-verification of Stanford study)
- themarkup.org/machine-learning/2023/08/14/ai-detection-tools-falsely-accuse-international-students-of-cheating
- Cross-confirms Stanford 61.3% finding.
- Weber-Wulff et al., 2023 — International Journal for Educational Integrity
- doi.org/10.1007/s40979-023-00146-z
- Cited as: A10 — peer-reviewed conclusion that AI text detectors are "neither accurate nor reliable." Used in pillars: 3, 4, 6.
Vendor documentation
- Turnitin — "Understanding false positives in Turnitin AI detection" (document level)
- turnitin.com/blog/understanding-false-positives-within-our-ai-writing-detection-capabilities
- Cited as: A3 — Turnitin's own document-level false-positive rate (<1%). Used in pillars: 3, 5, 6.
- Turnitin — "Understanding AI writing detection: False positive rates" (sentence level)
- turnitin.com/blog/understanding-the-false-positive-rate-for-sentences-of-our-ai-writing-detection-capability
- Cited as: A3 — Turnitin's own sentence-level false-positive rate (~4%). Used in pillars: 3, 5, 6.
- K-12 Dive — "Turnitin admits there are some cases of higher false positives in AI writing detection tool" (June 2023)
- k12dive.com/news/turnitin-false-positives-AI-detector/652221
- Cited as: A3 — Turnitin admission of higher-than-expected real-world rates.
- Inside Higher Ed — "Turnitin's AI detector: higher-than-expected false positives" (June 2023)
- insidehighered.com/news/quick-takes/2023/06/01/turnitins-ai-detector-higher-expected-false-positives
- Confirms Turnitin admission.
- GPTZero — "Falsely accused of AI cheating?" (vendor own guidance)
- gptzero.me/news/falsely-accused-of-ai-cheating
- Cited as: A9 — vendor's own statement that detection is "not definitive proof." Used in pillars: 1, 3, 4.
Named student cases
- Tech Learning — "He was falsely accused of using AI. Here's what he wishes his professor did instead" (William Quarterman, UC Davis)
- techlearning.com/news/he-was-falsely-accused-of-using-ai-heres-what-he-wishes-his-professor-did-instead
- Cited as: A5 — UC Davis Quarterman case cleared by Google Docs version history. A6 — MLK speech and Bible flagged by detectors. Used in pillars: 1, 2, 7.
- Rolling Stone — "Students Are Being Accused of Using AI" (Louise Stivers, UC Davis)
- rollingstone.com/culture/culture-features/student-accused-ai-cheating-turnitin-1234747351
- Cited as: A5 — UC Davis Stivers case, cleared by Google Docs version-history timestamps. Used in pillars: 1, 2, 7.
Lawsuits and public court records
- Detection Drama — "AI Detection Lawsuits: Every Student Case, Outcome, and What the Data Shows" (lawsuit tracker)
- detectiondrama.com/ai-detection-lawsuits
- Cited as: A4 — full lawsuit tracker. Newby v. Adelphi (won February 2026); Rignol v. Yale (pending); Doe v. UMichigan (pending); Yang v. UMN (dismissed without prejudice). Used in pillars: 1, 4, 8.
- GovTech — "Yale Student Suing Over Accusation of Improper AI Use"
- govtech.com/education/higher-ed/yale-student-suing-over-accusation-of-improper-ai-use
- Rignol v. Yale background.
- GovTech — "Student Sues University of Michigan Over AI Misconduct Accusation"
- govtech.com/education/higher-ed/student-sues-university-of-michigan-over-ai-misconduct-accusation
- Doe v. UMichigan background.
Academic frameworks and educator guidance
- LibreTexts — AI and College Writing: "How can I protect myself against false accusations of using AI?" (course material referencing Conrad's Blueprint)
- human.libretexts.org/Bookshelves/Composition/Specialized_Composition/AI_and_College_Writing
- Cited as: A7 — evidence-tools list; A8 — Kathryn Conrad's Blueprint for an AI Bill of Rights in Education. Used in pillars: 2, 4, 7.
Vendor shutdown announcement
- OpenAI — AI Classifier discontinuation (July 2023)
- Referenced in multiple news outlets and on OpenAI's blog. The Classifier was launched January 2023 and discontinued in July 2023, with OpenAI citing "low rate of accuracy."
- Cited as: A9 — vendor self-disclosure of inaccuracy. Used in pillars: 1, 3, 4, 6.
Free resources cited
- FIRE — Foundation for Individual Rights and Expression (United States)
- thefire.org
- ACLU — Student Speech and Privacy (United States)
- aclu.org/issues/free-speech/student-speech-and-privacy
- Office of the Independent Adjudicator (England and Wales)
- oiahe.org.uk
- Tertiary Education Quality and Standards Agency (Australia)
- teqsa.gov.au
- Citizens Advice (United Kingdom)
- citizensadvice.org.uk
- Pro Bono Students Canada
- probonostudents.ca
- Canadian Federation of Students
- cfs-fcee.ca
- Australian Human Rights Commission
- humanrights.gov.au
- International Center for Academic Integrity (ICAI)
- academicintegrity.org
Last reviewed: June 6, 2026. RE-VERIFY schedule: quarterly for vendor documentation; semi-annually for academic citations; on news of any new ruling for the lawsuit tracker.