Most of the noise around artificial intelligence at work focuses on the fear that a machine might eventually replace someone already on the payroll. But there is a much quieter, more pervasive shift happening at the very front door of the labor market: AI is increasingly deciding who even gets the chance to be hired in the first place.
A study from Stanford HAI highlights the scale of this gatekeeping. Researchers analyzed 3.4 million applicants across 150 employers and 11 different sectors, finding that approximately 90% of U.S. employers now use AI screening tools to sort and rank job seekers. The data reveals a troubling pattern: 26% of Black applicants and 15% of Asian applicants faced racial discrimination from these systems. The study points to an algorithmic monoculture, where a single vendor’s bias propagates across multiple employers simultaneously.
As these tools become the primary filter for the American workforce, the legal system is beginning to grapple with who is responsible when the math goes wrong. Two major lawsuits are currently testing the boundaries of accountability. In Mobley v. Workday, a court established an agency theory, suggesting that AI vendors can be held liable as agents of employers. Meanwhile, Kistler v. Eightfold AI takes a different approach, focusing on transparency.
While these cases move through the courts, the federal government has largely stepped back. The Equal Employment Opportunity Commission has shifted its focus, adopting a new National Enforcement Plan that moves away from disparate impact and toward intentional discrimination. The Biden-era AI guidance was removed in early 2025.
This has left a fragmented landscape at the state level. New York City’s Local Law 144 requires annual bias audits, though a December 2025 audit found enforcement ineffective. Colorado’s SB 24-205 was repealed and replaced by SB 26-189, which removed previous requirements for annual impact assessments.
The adoption of these tools is accelerating rapidly. As of August 2025, 57% of U.S. companies were using AI in their hiring processes, with 74% planning to increase that usage. Perhaps most significantly, 33% of employers say they expect AI to run their entire hiring process by the end of 2026.
The tension between these automated systems and the people they filter remains unresolved. While companies argue that human oversight remains the final word, the sheer volume of applications processed by these algorithms means that for many applicants, the machine’s initial ranking is the only one that matters.