When One Algorithm's Bias Follows Candidates Across Every Job Application

In 1970, a single fungal blight wiped out 15% of America's entire corn crop — not because the pathogen was especially powerful, but because 85% of the nation's corn shared the same genetic vulnerability. One flaw, one failure, everywhere at once. A landmark 2026 Stanford study found the same fragility is now built into corporate hiring. Researchers analyzed 4 million job applications across 156 employers and found that when a single vendor's AI screening tool carries a bias, that bias doesn't just affect one company — it follows candidates across every employer using the same system. They call it algorithmic monoculture. And more than 60% of Fortune 100 companies are participating in it right now. The numbers are striking: among applicants who applied to 10 positions through the same AI screening system, 4% were rejected everywhere — not because each company independently decided they weren't a fit, but because one algorithm made a correlated judgment that stuck. To reduce that systemic rejection risk below 0.1%, a candidate would need to apply to 25 or more positions. Compare that to just 10 if each employer's decisions were actually independent. For HR leaders, the episode breaks down what algorithmic monoculture actually means, why vendor-level audits are missing the real bias (hint: aggregate numbers mask position-level disparities), and what CHROs can do right now — from demanding position-by-position adverse impact analysis to building vendor diversification into procurement strategy.