Researchers and instructors warned that institutions’ growing reliance on AI “likelihood” detection tools is spawning a parallel marketplace. The concern is that some students can buy or access proprietary scores and then revise assignments until the “AI-likelihood” metric drops, undermining academic integrity procedures built on these outputs. The analysis describes a black-market pattern: students purchase reports tied to the same tools used by universities, edit essays to reduce flagged likelihood, and submit through the institution’s official integrity system. It also points to accuracy vulnerabilities, including elevated false positives for non-native English writers. The central operational risk for universities is twofold: false evidence creates due-process exposure, and monetized workaround access erodes the deterrence purpose of integrity systems. As AI detection becomes embedded in policy and discipline, the incentives for adversarial behavior rise. The story’s thrust is a call for universities to reassess how detection scores are interpreted, communicated, and governed—especially where the evidence standard may not match the statistical uncertainty of the tools.