What It Costs When a Machine Decides You Don’t Deserve a Chance

Erin Kistler has a Computer Science degree from Ohio State University. She has spent 19 years building expertise in product management and project management across government services, business, media, and entertainment. She has worked with AI. She has deployed data. She understands the technology. Since 2022, she has been applying to jobs at PayPal, Microsoft, and Netflix. She applied to two positions at PayPal in December 2025: a Senior Product Manager role on Venmo and a second Senior Product Manager role. She did not get an interview for either.

What Erin Kistler did not know, and what no one told her, is that the moment she submitted her application through a portal ending in “eightfold.ai/careers,” a machine had already begun building a dossier on her. It was pulling data from LinkedIn, GitHub, Stack Overflow, and other public sources. It was combining that with data about her location, her device activity, her browsing cookies. It was comparing her against the profiles of people who had been hired before her. It was predicting her future job titles. It was estimating how likely she was to “join a company like the hiring company.” It was looking at the hiring managers and high-performing employees at PayPal and asking how closely she resembled them. Then it assigned her a number between 0 and 5.

No human being ever saw her resume. The number was too low.

Sruti Bhaumik holds a Bachelor of Arts in Chemistry from Bryn Mawr College and a Master of Science from the University of Pittsburgh. She has over a decade of experience in project management, with a specialty in improving digital product experiences using data and AI. She lives in Walnut Creek, California. She applied to Microsoft for a position called Senior Technical Program Manager for Responsible AI. The job title alone tells you she was qualified to question whether AI systems were being used responsibly. She received an automated rejection two days after applying. She was never interviewed. She applied again in December 2025 for a second Microsoft program manager role. Same result.

Think about what that automated rejection email means when you are sitting on the other end of it. You spent time on that application. You tailored your resume. You researched the company. You maybe felt a flicker of hope. Then two days later, a form letter arrives. No explanation. No human name attached. No indication that your 10 years of experience was even read. Because it was not. An algorithm compared you to a billion other people and decided you did not score high enough. And the score, the data used to generate it, and the sources Eightfold pulled from are all invisible to you. You cannot see them. You cannot challenge them. You cannot correct a factual error even if one existed. You have no legal right to any of it, under the current Eightfold system.

These are not edge cases. The complaint alleges that nearly two-thirds of large companies (those with more than 5,000 employees) now use AI technology like Eightfold’s to pre-screen candidates. Thirty-eight percent use it to match and rank them. Eightfold alone claims to screen millions of candidates for over 100 corporate clients. Every single one of those candidates was subjected to surveillance and scoring they never consented to. Every single one of them may have data errors in their profiles that got them silently thrown out. None of them were told. Most of them still do not know.

The laws that are supposed to protect these people have existed since 1970. Congress wrote the Fair Credit Reporting Act specifically because lawmakers were terrified of exactly this: “impersonal blips and keypunch holes in a stolid and unthinking machine” that could “ruin his reputation without cause, and make him unemployable or uninsurable.” That was 1970. They saw it coming. And Eightfold, a well-funded AI company headquartered in Santa Clara with access to lawyers and full knowledge of published regulatory guidance, chose not to comply anyway.

Verbatim: What the Lawsuit Actually Says

The following quotes are taken directly from the complaint filed January 20, 2026 in the Superior Court of California, Contra Costa County, by Outten & Golden LLP and Towards Justice on behalf of plaintiffs Erin Kistler and Sruti Bhaumik.

“Eightfold’s technology lurks in the background of job applications for thousands of applicants who may not even know Eightfold exists, let alone that Eightfold is collecting personal data, such as social media profiles, location data, internet and device activity, cookies and other tracking, to create a profile about the candidate’s behavior, attitudes, intelligence, aptitudes and other characteristics that applicants never included in their job application.” Class Action Complaint, Paragraph 1, Kistler & Bhaumik v. Eightfold AI Inc. (2026)

“Lower-ranked candidates are often discarded before a human being ever looks at their application.” Class Action Complaint, Paragraph 6

“Eightfold’s policy also makes clear that the ‘personal data’ it collects and uses for evaluating employment includes ‘[i]nferences drawn from any of [the various sources of information it collects] to create a profile about [the job applicant] reflecting the [applicant’s] preferences, characteristics, predispositions, behavior, attitudes, intelligence, abilities, and aptitudes.'” Class Action Complaint, Paragraph 66, citing Eightfold’s own privacy policy

“Eightfold trains the Match Score model ‘specifically for candidate-job fit using supervised learning on hiring data’ that is external to the specific candidate’s job application.” Class Action Complaint, Paragraph 68, citing Eightfold’s own technical paper (arxiv.org/pdf/2507.02087v1)

“With a growing list of over 100 customers—including Microsoft, Morgan Stanley, Starbucks, BNY, Paypal, Chevron and Bayer—Eightfold boasts ‘unparalleled’ capabilities that ‘screen[] millions’ of candidates.” Class Action Complaint, Paragraph 8

“Eightfold voluntarily ran a risk of violating the law substantially greater than the risk associated with a reading that was merely careless.” Class Action Complaint, Paragraph 136(e)

The Scale of Harm: Who Loses When AI Gatekeeps Employment

Public Health

Unemployment and job rejection at scale carry documented mental and physical health consequences. The secrecy of Eightfold’s system compounds those harms in specific ways.

• Job rejection without explanation prevents applicants from understanding what went wrong or improving future applications. The psychological harm of repeated unexplained rejections is documented and severe: it erodes self-worth, increases anxiety, and can trigger or worsen depression. When the rejection is produced by an invisible algorithm using data the applicant never submitted, there is no constructive feedback loop available.
• Applicants who are incorrectly scored due to data errors have no mechanism to identify or correct those errors. A resume that was incorrectly parsed, a social media profile that does not belong to them, or an inference the AI drew incorrectly about their “predispositions” can follow them from application to application across all 100+ Eightfold client companies, compounding unemployment for an indeterminate period.
• Extended unemployment directly increases risks of physical health deterioration, loss of health insurance coverage, housing instability, and food insecurity. The complaint notes that both named plaintiffs have been applying since 2022 and 2023 respectively, suggesting multi-year exposure to this system’s effects.
• The CFPB guidance cited in the complaint explicitly recognizes that these harms are not speculative. Congress identified them in 1970 and strengthened protections specifically because of them in 1996. The documented public health dimension of employment discrimination via biased or inaccurate records is the reason these statutes exist.

Economic Inequality

AI hiring systems trained on historical data replicate and amplify existing economic inequalities at industrial scale. Eightfold’s design makes this risk structural, not incidental.

• The Match Score algorithm compares applicants against “ideal candidates” and high-performing employees at the hiring company. If those reference employees reflect prior hiring biases (which research consistently shows they do), the algorithm systematically disadvantages anyone who does not fit that historical mold: women, people of color, people from non-elite schools, career changers, workers with employment gaps, and people from lower-income zip codes.
• “Title progression and seniority fit” scoring includes a calculation of the “similarity in look and feel between the hiring company and the candidate’s past companies.” This is a direct economic inequality multiplier. Candidates who have worked at smaller companies, non-profits, government agencies, or in industries not well-represented in Eightfold’s database will be systematically downscored compared to candidates with blue-chip employer histories.
• The “hireability” factor, defined as “the likelihood that the job candidate will join a company like the hiring company,” encodes economic geography and class signaling into the score. A candidate from a working-class background, a historically Black college, or a rural region may present signals the algorithm reads as low hireability, with no opportunity to correct that inference.
• Eightfold’s system is used disproportionately by the largest companies (the complaint notes nearly two-thirds of companies with more than 5,000 employees use AI screening). These are precisely the employers who offer the best wages, benefits, and career advancement. Being systematically filtered out by an AI at this tier of the labor market forecloses access to economic mobility at the highest-stakes level.
• The complaint notes that employers using Eightfold include Chevron and Bayer, industries with historically documented racial and gender disparities in hiring. An AI trained on those companies’ historical hiring data carries those disparities forward algorithmically, at scale, without any human ever noticing or being held accountable.
• Statutory damages under FCRA range from $100 to $1,000 per violation. While this sounds modest per person, the complaint covers millions of applicants. The aggregate exposure to Eightfold and its employer-clients could be enormous. Those damages, however, do not undo years of lost wages for applicants who were quietly filtered out.

What Eightfold’s Violations Are Worth — Per Person

The Watchlist, Your Rights, and How to Fight Back

This lawsuit is a beginning, not a resolution. Here is who has power over what happens next, and what you can do right now.

Corporate Leadership Under Legal Scrutiny

The complaint names Eightfold AI Inc. as defendant. The following corporate clients are named in the complaint as employers who used Eightfold’s system without the required FCRA certifications or applicant disclosures:

• PayPal — named as an employer that used Eightfold’s Evaluation Tools to process plaintiff Erin Kistler’s December 2025 applications without FCRA certification or disclosure.
• Microsoft — named as an employer that used Eightfold’s Evaluation Tools to process plaintiff Sruti Bhaumik’s July 2025 and December 2025 applications without FCRA certification or disclosure.
• Morgan Stanley, Starbucks, BNY, Chevron, and Bayer are named in the complaint as Eightfold clients, though they are not named as defendants in this action.

Regulatory Watchlist

These agencies have jurisdiction over the conduct described in this complaint. Contact them. File a complaint. Make noise.

• CFPB (Consumer Financial Protection Bureau) — The primary federal agency that administers the FCRA. The CFPB published explicit 2024 guidance stating that AI hiring tools are covered by the FCRA. It has authority to investigate, fine, and issue injunctive orders against consumer reporting agencies. File a complaint at consumerfinance.gov.
• FTC (Federal Trade Commission) — Joint enforcement authority over the FCRA alongside the CFPB. The FTC explicitly stated in 2013 that employment background screening apps are covered by the FCRA and that claiming otherwise is not a defense. Contact at ftc.gov.
• EEOC (Equal Employment Opportunity Commission) — Relevant if AI scoring patterns reveal disparate impact discrimination on the basis of race, sex, national origin, or age. The complaint notes Eightfold must certify its reports are not used in violation of equal employment opportunity laws — certifications it allegedly never obtained.
• California Department of Justice (Consumer Protection) — Has enforcement authority over the ICRAA and UCL violations alleged in this complaint. Contact at oag.ca.gov.
• California Privacy Protection Agency (CPPA) — Enforces the California Consumer Privacy Act (CCPA), which has provisions overlapping with the data collection practices described here.

Immediate Action Steps

• If you have applied to a job through a portal with “eightfold.ai” in the URL, you may be a class member in this lawsuit. Track the case: Kistler and Bhaumik v. Eightfold AI Inc., Superior Court of California, Contra Costa County, filed January 20, 2026. Follow classaction.org for updates on the case status.
• Contact the plaintiffs’ legal teams directly if you believe you were subjected to Eightfold’s Evaluation Tools without disclosure. Outten & Golden LLP (jsagafi@outtengolden.com) and Towards Justice (rachel@towardsjustice.org) are counsel for the class.
• Check every job application portal URL you submit through. If you see “eightfold.ai” in the address bar, you are being processed by this system. Document it with a screenshot including the URL and the date.
• File a CFPB complaint against Eightfold AI Inc. at consumerfinance.gov/complaint. Volume of complaints triggers regulatory investigation priority. A single complaint is data; ten thousand complaints is political pressure.
• Share this story with your union, your mutual aid network, your workplace organizing group. AI hiring discrimination is a collective labor issue, not an individual consumer problem. The workers most harmed by opaque AI screening are the ones who can least afford lawyers.
• If you are a researcher, journalist, or policy advocate: the CFPB’s 2024 guidance on AI hiring tools is a public document (linked in this article’s sources). Push your representatives to codify it into statute before a future administration can retract it.
• Support local worker centers and job seeker advocacy organizations in your area. They are already doing the ground-level work of helping people navigate opaque hiring systems. They need resources, volunteers, and visibility.