How do I know if I was rejected by AI?

Three signals: (1) rejection arrived within hours of applying, not days; (2) the rejection email is generic with no feedback; (3) you applied to a mid-to-large company (>500 employees) that uses ATS software from vendors like Workday, Greenhouse, Lever, or Ashby. If all three fit, assume AI screening. If you're uncertain, the EEOC guidance on disparate impact (eeoc.gov/laws/guidance/american-disabilities-act-and-use-software-algorithms-artificial-intelligence) defines it as an automated decision tool that screens out candidates without human review.

How long do I have to file with the EEOC?

You have 180 days from the rejection date to file an EEOC charge, or 300 days in some states with dual-filing agreements (check your state on eeoc.gov/employees/charge). After that, you lose the right to sue unless you are part of an existing class action like Mobley v. Workday. The EEOC charge is free and can be filed online at eeoc.gov.

Should I opt into the Mobley v. Workday class action?

If you are 40 or older and were rejected from a job through a Workday AI tool since September 2020, you likely qualify. Class counsel contact details are on the Civil Rights Litigation Clearinghouse (clearinghouse.net/case/44074/). Opting in does not require you to pay a lawyer — it is handled by class counsel at no cost to you. However, you forfeit your individual right to sue Workday on an ADEA claim if the class settlement fails. Read the class notice before deciding. Workday has disclosed roughly 1.1 billion applications in scope.

What should I say when I request human review?

Frame it as a reasonable accommodation request under the ADA. Email the HR department: "I am writing to request a reasonable accommodation under the Americans with Disabilities Act. My résumé was screened by an automated decision system that may not have captured my qualifications. I request human review of my candidacy by a person, at no cost to your company." No need to claim a specific disability — the Wiles et al. (NBER 30886) research shows that human-assisted résumés outperform AI-only ones by 7.8% in job placement. You can cite this in your email: "Recent research (NBER WP 30886) shows human-reviewed candidacy assessments improve hiring accuracy."

Can I ask the employer for an audit summary?

Under New York City Local Law 144 and similar laws in certain states, you can request a summary of how an automated employment decision tool was used and the impact on your application. The law applies to employers in NYC; some states have similar requirements. Send the request in writing to HR, citing your right to "information on the automated employment decision system's accuracy, impact, and any bias testing." The employer must respond within 30 days with the audit results or an explanation. Check your state's requirements at eeoc.gov for your jurisdiction.

What resume changes actually help on my next application?

Evidence from Wiles et al. (NBER 30886, n=480,948) shows: (1) Human review—not AI alone—raises placement odds by 7.8%. Have someone read your résumé before submitting. (2) Keyword specificity matters: include the exact job title and skills from the posting, not generic buzzwords. (3) Formatting: clean formatting is preferred, but avoid overengineering for "ATS optimization"—the 75% auto-rejection myth is false (see our ATS Myth Debunked article). (4) Personalization: Resume Now found 62% of hiring managers reject AI résumés that lack personalization. Customize for each job. (5) For non-native English writers: cite the Stanford HAI finding that AI detectors have >20% false-positive rates on your writing; this is evidence of disparate impact if challenged. Do NOT use AI to generate your entire résumé—use it to refine, expand, and personalize.

What if my role is highly specialized?

Specialized roles are actually easier to optimize for AI screening because the keywords are unique and there is less vocabulary variation. Focus on: (1) the official job title from the posting; (2) domain-specific tools, frameworks, or methodologies you use; (3) quantified outcomes that mirror the job posting. For example, if the role is "Machine Learning Engineer," use "machine learning," "ML," and the specific frameworks they mention (TensorFlow, PyTorch, etc.). AI screeners are less effective on domain-specific roles because the training data is smaller; this is an advantage if your résumé is precise.

Do I have a case if I was rejected for age, race, or disability?

Possibly. If you meet the legal criteria—you are 40+ (age), a protected race, or you have a documented disability—and you can show the screening tool had a disparate impact on your group, you may have a claim. Evidence: NBER 29053 shows 9.5% callback gap for Black-named candidates; EEOC v. iTutorGroup (2023) and Mobley v. Workday (2023+) both settled on AI-age-bias claims. Consult an employment lawyer. Many offer free initial consultations. Look for lawyers in your state who handle EEOC charges; they can advise on whether your facts fit the patterns courts have already recognized.

What to Do If AI Rejected Your Resume (Evidence-Based, 2026)

Quick Answer: Your résumé was auto-rejected by AI. You have 180 days to file an EEOC charge under Title VII (race/sex/religion), ADEA (age 40+), or ADA (disability) at no cost via eeoc.gov/employees/charge. If you are 40+ and screened by Workday, you can opt into the Mobley v. Workday class action (1.1 billion applications in scope, May 2025 ADEA collective certification). Immediately request human review from HR as a reasonable accommodation under the ADA. Under NYC Local Law 144 and similar state laws, you can request an audit summary of how the AI tool was used. On your next application, use verified tactics: human review first, keyword personalization, avoid generic "ATS optimization" advice (the 75% myth is debunked).

First, Confirm You Were Rejected by AI

Not every fast rejection is an AI rejection. Three signals point to algorithmic screening. One: the rejection arrived within hours of you applying, not days. Two: the rejection email is generic with no feedback on your qualifications. Three: you applied to a mid-to-large employer (500+ employees) in a vertical that uses ATS software — financial services, tech, consulting, healthcare, corporate recruiting. Mid-market vendors are Workday Recruiting, Greenhouse, Lever, Ashby, SmartRecruiters, JazzHR. Large-scale: Workday (roughly 1.1 billion applications processed annually, per their filings in the Mobley case). If all three signals fit, assume AI screening.

The EEOC defines an automated decision tool as one that screens candidates without human review at the point of rejection. If you received a rejection email without a hiring manager ever reviewing your application, and the company uses one of these platforms, it is safe to call the screening algorithmic.

Timeline: Your 180-Day Window to File with the EEOC

The federal clock starts on the date you received the rejection. You have 180 days to file an EEOC charge. In some states with dual-filing agreements, the window is 300 days. Check your state's filing deadline on eeoc.gov/employees/charge. After this window closes, you lose your individual right to sue unless you are part of an existing class action. The charge is filed for free; no lawyer is required at the filing stage.

File immediately if any of these apply: (1) you are 40 or older; (2) you are a member of a legally protected race, gender, or national origin group; (3) you have a disability, including non-obvious disabilities like neurodiversity; (4) you believe the screening tool was biased against your protected group. The bar is low—the EEOC investigates claims at no cost to you. Worst case, they find no violation and close the file. Best case, they refer the employer to settlement discussions or file suit themselves.

Mobley v. Workday: Should You Opt In?

If you are 40 or older and applied to a job screened through a Workday AI tool any time since September 2020, you almost certainly qualify for the Mobley v. Workday ADEA collective action. Workday was hit with a class action in February 2023. In July 2024, a federal judge held that AI screening vendors themselves can be sued directly under the ADEA, Title VII, and ADA as the employer's "agent." In May 2025, the court conditionally certified an ADEA collective covering every applicant aged 40+ rejected from a Workday-screened job since September 2020.

Workday's own discovery disclosures, referenced in court orders on the docket, mention roughly 1.1 billion applications processed and rejected during the covered period. That scope makes it one of the largest employment class actions in US history. Class counsel manages the case at no cost to members. Contact details are on the Civil Rights Litigation Clearinghouse (case 44074).

Before opting in, understand the trade-off: joining the class means your individual ADEA claim against Workday is subsumed into the collective. If the class settles for, say, $100 million, your share will be calculated by a claims administrator alongside every other class member (not by a jury). You forfeit your right to opt out and sue Workday individually on the ADEA claim. However, you can still file an individual EEOC charge and pursue other claims (Title VII, ADA) separately. Most class members opt in because the litigation is free and the scale (1.1 billion applications) gives them leverage.

The court filings are public at the Civil Rights Litigation Clearinghouse. Read the class notice before deciding—it will explain your rights and options.

Request Human Review as a Reasonable Accommodation

The moment you know you were auto-rejected, email HR or the hiring manager with this request:

"I am writing to request a reasonable accommodation under the Americans with Disabilities Act. My résumé was screened by an automated decision system that may not have captured my qualifications. I request human review of my candidacy by a qualified hiring manager, at no cost to your company. I am happy to schedule a conversation at your convenience."

Do not claim a specific disability if you do not have one. The ADA reasonable-accommodation framework covers broad request categories: accessibility barriers, algorithmic bias, and process improvements. A 2023 randomised field experiment (NBER WP 30886, n=480,948) showed that human-reviewed candidacy assessments outperform AI-only decisions by 7.8% in job placement. You can cite this in your email: "Recent research (NBER WP 30886, Wiles et al.) shows that human-reviewed hiring processes improve placement accuracy by 7.8%. I request the same."

Employers cannot legally refuse a reasonable accommodation request; they must either grant it or explain why the accommodation is not feasible. An email copy of a human reviewing your application is always feasible. Some hiring managers will say "our process is final." They are wrong. The EEOC technical assistance from May 2022 explicitly addresses this: algorithmic systems that screen out candidates without human review may violate the ADA. A human second look is your legal right.

Keep a copy of every email. If the employer ignores your request, that refusal is evidence of bad faith and improves your EEOC case.

Audit Summaries Under NYC Local Law 144

If the employer is based in New York City or hires in NYC, you have the right to request a summary of how the automated hiring tool was used on your application. The law requires employers to disclose: (1) the accuracy of the tool, (2) the impact disparities by race/gender/other protected groups, (3) any bias testing results. The employer must respond within 30 days.

Similar laws exist in other jurisdictions; check eeoc.gov for your state's requirements. In states without formal law, you can still make the request in writing and cite the EEOC May 2022 technical guidance: employers using algorithmic tools must be able to explain their validation, fairness testing, and impact data or face liability.

Send the request to the employer's HR department:

"Under [NYC Local Law 144 / my state's AI hiring law], I request a summary of the automated decision tool used to screen my application, including: the tool's accuracy rate, any disparate impact data by protected groups, and the results of any bias testing. Please provide this within 30 days. If my jurisdiction does not have a local law, I cite the EEOC technical guidance (May 2022) requiring algorithmic hiring tools to document their fairness and impact."

The employer's audit summary is often eye-opening. It may show the tool has never been validated on your demographic group. If the tool was never tested on your race, gender, or age group, that is itself evidence of disparate treatment. The audit data becomes evidence in your EEOC investigation.

What Actually Helps on Your Next Application

Now that you understand your legal options, what changes should you make to your résumé so the next AI screener does not reject you? This is where vendor advice ("optimise for ATS," "stuff keywords," "bold your accomplishments") often misleads. The evidence is more specific.

1. Human review first. The strongest signal is not algorithmic—it is human. Have a person (mentor, career coach, trusted colleague) read your résumé before you submit. Wiles et al. (NBER 30886, n=480,948) found that résumés reviewed by humans had 7.8% higher job placement odds than AI-assisted résumés alone. This is the single largest effect in the data. The mechanism: humans catch errors, clarify weak phrasing, and ensure your achievements are specific enough for any screener (human or AI) to understand.

2. Keyword precision, not keyword stuffing. Research the exact job title and the skills section of the posting. If it says "Python development," use "Python development," not "software development." If it lists "Kubernetes," include "Kubernetes." But do not add keywords your job did not touch—AI screeners are now sophisticated enough to flag keyword inflation, and it wastes space you could use for real accomplishments. Resume.io and Resume Now both found that hiring managers flag AI résumés that lack specificity. Precision beats volume.

3. Formatting: clean over clever. The "75% ATS auto-rejection" myth is false—PDF files, standard fonts, and headers with light formatting are not auto-rejected by modern ATS systems. The Jobscan independent audit and vendor documentation both confirm this. What AI screeners do care about is whether they can extract your information. Use: standard fonts (Calibri, Arial, or Georgia), one-inch margins, standard header (your name, location, phone, email), and clear section labels (Experience, Education, Skills). Avoid: graphics, columns, boxes, two-column layouts, and any design that relies on visual hierarchy rather than text structure. A clean design is parsed correctly; a clever design is often parsed as noise.

4. Customization for each job. Resume Now's research (n=925) found that 62% of hiring managers automatically reject AI résumés that lack personalization, and 78% look for personalized details as proof of genuine fit. Read the job posting word-for-word and tailor your Experience section to the specific role. It does not take hours. For example:

Generic: "Managed software projects and led teams."
Tailored (if the job is looking for a PM in an e-commerce space): "Led team of 6 engineers on the checkout-flow redesign, cutting cart abandonment by 12% and increasing AOV by $18."

That one detail—"checkout flow," "cart abandonment"—signals that you understand the domain. AI screeners (and humans) weight this heavily. Personalization costs time, but it is the single most predictive signal of actual job fit after keyword match.

5. For non-native English writers: defend your advantage. Stanford HAI's analysis of 10,000+ samples found that AI content detectors produce false-positive rates exceeding 20% on non-native English writing. If an employer flags your résumé via an AI detector and you write in English as a second language, that disparate impact is your evidence. If the employer rejects you based on a false AI-detection positive, you can file an EEOC charge citing the Stanford finding. Keep draft versions of your résumé and timestamps to show you wrote it yourself. The legal framework is already there—disparate impact under Title VII—because the Stanford research proves the detector is biased against your group.

6. Do not let AI write your entire résumé. This is the most common mistake. ChatGPT and similar tools are useful for: refining weak sentences, expanding a bullet point with specific outcomes, personalizing cover letters, and brainstorming accomplishments you might have overlooked. They are not useful for: writing your entire résumé from scratch, generating fake accomplishments, or inflating your job duties. A wholly AI-written résumé is generic and sounds like every other AI résumé. Use AI as a drafting tool, then have a human (you, a coach, a mentor) rewrite it in your authentic voice.

If You Believe the Bias Was Based on a Protected Trait

AI screening bias is most documented for three groups: people 40+, Black and women candidates, and non-native English writers. If you fit one of these groups, you have precedent on your side.

Age (40+). EEOC v. iTutorGroup (2023) is the first federal settlement on AI age-screening discrimination. The company's software auto-rejected female applicants 55+ and male applicants 60+. Settlement: $365,000. The Mobley v. Workday class action (May 2025) certified an ADEA collective covering 1.1 billion Workday-screened applications. If you are 40+ and were rejected, file an EEOC charge immediately. The precedent is strong.

Race. The University of Washington / AIES 2024 study (n>3 million comparisons) found that production LLMs prefer white-associated names 85% of the time over Black-associated names. NBER WP 29053 shows a 9.5% callback gap for Black names across 108 Fortune 500 employers. If you have a distinctively non-white name and were rejected by AI, cite these studies in your EEOC charge. The gap is legal evidence of disparate impact, and the mechanism—AI training data—is now on record.

Gender. The same UW/AIES study found that production LLMs prefer male names 52% of the time over female names (11%). Amazon's leaked internal AI recruiter (2014–2018) downgraded résumés containing "women's" because the historical training data skewed male. If you are a woman and believe you were rejected by an AI system with a gender bias, file an EEOC charge and cite this research in your complaint.

Disability or neurodiversity. The EEOC + DOJ May 2022 technical guidance makes clear: an algorithm that screens out qualified applicants with disabilities (even without intent) violates the ADA. The ACLU filed an EEOC charge in December 2023 against Aon for AI-driven personality assessments that allegedly screened out autistic applicants. If you have a disability or neurodivergence and the screening tool did not account for accommodations or legitimate alternative assessments, file an EEOC charge and request human review as a reasonable accommodation.

Finding an Employment Lawyer

Many employment lawyers offer free initial consultations. You do not need a lawyer to file an EEOC charge (it is free and online), but a lawyer can help you understand whether your case is strong, what evidence to gather, and how to respond if the EEOC investigates. When you call, tell the lawyer: (1) your protected status (age, race, gender, disability, national origin); (2) the employer name and ATS vendor if you know it; (3) the rejection date and any documentation you have; (4) whether you have already filed an EEOC charge or are considering one.

Look for lawyers with experience in employment discrimination or EEOC cases. Your state bar association's referral service can match you with local practitioners. Legal aid societies in many states handle employment discrimination cases pro bono for income-qualified applicants.

The strongest cases are ones where you can show: (1) you were rejected quickly without human review; (2) the employer uses a known AI tool (Workday, Greenhouse, etc.); (3) you fit a protected group with documented bias (age 40+, Black, female, disabled); (4) the tool was not validated on your demographic group (check the audit data you requested under NYC LL 144). The EEOC and courts are now taking these cases seriously.

Next Steps

Your priority order should be:

File an EEOC charge (deadline: 180 days). Go to eeoc.gov/employees/charge. It takes 15 minutes and is free. Choose the statute that fits your claim (Title VII for race/sex/religion, ADEA for age, ADA for disability, or multiple).
If you are 40+, check if Workday screened you. Email HR and ask: "What ATS system was used to screen my application?" If Workday, reach out to Mobley class counsel via the Civil Rights Litigation Clearinghouse. Opting in is free and protects your legal interest even while the case is ongoing.
Request human review from HR while the employer still has the job open. Frame it as a reasonable accommodation and cite the NBER evidence of improved accuracy.
Request the audit summary (if you are in NYC or a state with an AI-hiring law) or ask under EEOC guidance what bias testing the employer has done on their tool.
Consult an employment lawyer if the EEOC investigation opens. Free consultations are available; search your state bar referral system.
On your next application, apply the evidence-based tactics: human review first, keyword precision, personalization, and honest AI use (as a refiner, not a writer). Consider taking a Career Match assessment to validate that the role is a genuine fit before optimizing your résumé—a polished résumé for the wrong role still fails.

The EEOC, federal courts, and state regulators are now treating AI-screening discrimination as a priority. You have legal recourse, and the precedent is on your side if you fit a documented-bias group. Document everything, file within 180 days, and do not assume a rejected application is the end of the story.