The score draws on a database built from the resumes and social profiles of more than a billion job seekers, and it can push some applicants ahead of others when interviews are handed out.
Erin Kistler has brought a class action against Eightfold AI, a Silicon Valley firm whose recruiting software hundreds of employers rely on — among them some of the places she sought work. The case was filed in January in a California court.
Kistler has worked in product management for close to two decades and considers herself qualified for every job she went after. Over the past four years she has sent out thousands of applications and has never once been called in for an interview.
What the software returns on each candidate is a rating from 0 to 5 — a forecast of how well that person would perform in a particular role, drawn from a database Eightfold keeps in constant update. Kistler's lawyers say the case ranks among the first to argue that automated screening of this kind amounts to a hidden consumer report or candidate file — one that grades applicants on how likely they are to succeed while leaving them no way to view or contest what it concludes.
A spokesperson for Eightfold wrote to the Guardian in an email that the company considers the claims without merit and plans to mount a vigorous defense.
Rachel Dempsey, a lawyer acting for Kistler, argues that job seekers should get the same visibility into what feeds hiring decisions that Americans already have into their credit files, which would let them challenge errors or address whatever is weighing against them. Without that view of the algorithms, she said, there is no way of knowing whether they openly discriminate against people applying for jobs.
Kistler's suit is one of a number of recent court fights over how employers use AI in decisions about workers. Employees have taken Meta to court over an in-house AI system they say singled them out for job cuts on the basis of parental or medical leave they had taken; the Guardian reported that Meta had nothing to say about that case when asked for comment. IBM faces a recently filed case claiming its AI tools worked against workers on the basis of age; the company said it does not use AI to screen candidates out automatically and that it neither condones nor tolerates discrimination of any sort.
Rules requiring employers to say anything reach only so far. A New York City law in force since 2023 obliges employers using automated hiring systems to run bias audits every year and to notify candidates in advance, but it covers only software that substantially assists the decision or takes it over, leaving a gap wherever humans remain part of the process. Illinois and Colorado have recently passed laws barring employers from using AI tools that lead to unlawful discrimination.
Beyond such measures there is no across-the-board requirement to give notice, says Ifeoma Ajunwa, a law professor at Emory University who founded the school's program on AI and the future of work — so workers are not always told when a machine is assessing them.