Home › Case studies
Case studies.
Ten public failures of commercial AI detection — and the signed proof we ship against each one. Every claim on this page has a verifier, a try-it endpoint, or a bounty.
Case 01
Equity crisis
Turnitin flags non-native English writers as AI at 61%.
The failure that got Turnitin removed from a dozen universities.
Case 02
The adversarial hole
Every academic AI detector collapses under adversarial paraphrase.
The one attack that breaks GPTZero, Originality.ai, Binoculars, and FastDetectGPT — and where we score 99.6%.
Case 03
Benchmark saturation
The RAID leaderboard’s 99% club fit on the public labels.
Why the top 10 detectors all hit 99% — and why that number is meaningless.
Case 04
Silent updates
GPTZero, Grammarly, and Trinka silently update. Their scores are not reproducible.
Same text, different day, different score — evidence value goes to zero.
Case 05
Court admissibility
Every commercial AI detector fails Daubert. Ours doesn’t.
Federal Rule of Evidence 702 requires known error rate + reproducibility. Only signed, deterministic detectors clear that bar.
Case 06
Language coverage
None of the top 10 leaderboard detectors submitted non-English results.
Every top-10 detector was calibrated on English. The multilingual world was left out.
Case 07
No AI. No GPU. No data center. No power draw.
Detecting AI text should not require a data center.
Every competitor runs on GPUs or hosted APIs. We run on any CPU.
Case 08
FERPA · GDPR · EU AI Act
Every SaaS detector transmits student text to their servers.
The compliance wall that blocks Turnitin, GPTZero, Grammarly at K-12 and EU institutions.
Case 09
Industry-wide
AI benchmarks are saturated. Real measurement moved elsewhere.
MMLU 99%. SWE-bench dropped. RAID top-10 is the same story.
Case 10
The kill shot
Here’s your API key. Try to defeat our detector.
Bring your 100 toughest texts. If we fail more than 5%, refund plus we publish the failure.