The iTutorGroup AI Age-Discrimination Settlement
The iTutorGroup EEOC settlement is a 2023 agreement in which the online tutoring company iTutorGroup paid $365,000 to resolve a U.S. Equal Employment Opportunity Commission lawsuit alleging that its recruiting software automatically rejected applicants based on age. According to the EEOC, the company's hiring tool was programmed to reject women applicants aged 55 or older and men aged 60 or older, a violation of the Age Discrimination in Employment Act (ADEA). It was the EEOC's first settlement of a case centered on AI-driven hiring discrimination.
What was the iTutorGroup EEOC settlement?
iTutorGroup operated a set of brands that recruited tutors, many of them U.S.-based, to teach English to students in China. The EEOC alleged that the company's online application software was configured to auto-reject candidates by date of birth: female applicants 55 and older, and male applicants 60 and older, were screened out before any human reviewed their qualifications.
The case became public when one applicant, after her real birth year drew a rejection, resubmitted the same application with a more recent date of birth and received an interview offer. That contrast supplied direct evidence that the rejection turned on age, not qualifications.
The EEOC filed suit in May 2022 in the U.S. District Court for the Eastern District of New York (EEOC v. iTutorGroup, Inc., No. 1:22-cv-02565). In August 2023 the parties entered a consent decree settling the matter. Key terms included:
A monetary payment of $365,000 to a group of more than 200 rejected applicants.
A requirement that iTutorGroup adopt anti-discrimination policies and conduct training.
Provisions for the company to consider for hire qualified applicants who had been screened out.
An injunction against age- and sex-based hiring practices, with EEOC oversight during the decree period.
The settlement did not require a court finding of liability, which is typical for a consent decree. The legal significance is the precedent: a federal agency treated an automated employment decision tool (AEDT) the same way it would treat a human manager who refused to hire older workers.
Why does the iTutorGroup case matter for AI hiring?
The case matters because it set the EEOC's enforcement posture on a single principle: the law that governs human hiring decisions governs algorithmic ones too. There is no AI exemption in the ADEA, Title VII, or the Americans with Disabilities Act (ADA).
Three points make iTutorGroup a reference case for employers and vendors.
The rule did not need to be subtle to be illegal. The iTutorGroup tool used an explicit age cutoff, a clear hard-coded rule. But the EEOC's broader guidance treats both deliberate cutoffs and statistically biased outcomes as actionable, which extends the lesson well past simple date-of-birth filters.
The employer, not the software vendor, was held responsible. iTutorGroup paid the settlement. Liability for discriminatory hiring sits with the entity making the employment decision, regardless of who built the tool. A vendor's assurance that a product is "bias-free" does not transfer that liability.
A single contradictory data point exposed the system. The reapplication produced a before-and-after comparison the company could not explain away. Employers running opaque screening tools often cannot say why a given candidate was rejected, which is itself a legal exposure.
For a fuller treatment of how the agency frames algorithmic hiring under federal civil rights law, see The AI Table's analysis of EEOC technical guidance on AI hiring.
How does iTutorGroup fit the EEOC's AI initiative?
iTutorGroup did not appear in isolation. In late 2021 the EEOC launched an agency-wide effort to examine how artificial intelligence and other tools are used in employment decisions, with the stated goal of ensuring those tools comply with the civil rights laws the agency enforces. The initiative produced two technical assistance documents: one in May 2022 on the Americans with Disabilities Act and algorithmic screening, and one in May 2023 on Title VII and the use of software, algorithms, and AI in selection procedures.
The iTutorGroup settlement landed in August 2023, three months after the Title VII guidance, and gave the initiative its first enforcement result. The guidance told employers how the agency reads the law; the settlement showed it would litigate. The EEOC's strategic enforcement plan for 2024 through 2028 also lists technology-driven discrimination as a priority, which suggests iTutorGroup will be followed by further cases rather than treated as an isolated event.
How does the ADEA apply to automated screening tools?
The Age Discrimination in Employment Act of 1967 protects workers age 40 and older from employment discrimination based on age. It applies to employers with 20 or more employees and covers hiring, firing, pay, promotion, and other terms of employment.
An automated screening tool does not change the analysis. If a tool produces a hiring outcome that disfavors workers 40 and older, the ADEA reaches it under two theories:
Disparate treatment: intentional discrimination, such as a rule that rejects applicants over a set age. The iTutorGroup cutoff is a clear example.
Disparate impact: a facially neutral practice that disproportionately harms a protected group. A resume screener that down-ranks gaps in employment, older graduation dates, or certain phrasing can produce age-skewed results even with no explicit age field.
The disparate-impact theory is the harder problem for AI screening. A model trained on a company's past "successful" hires reproduces the age distribution of those hires. If the historical workforce skewed young, the model learns to prefer markers of youth, and the output discriminates even though the age field never appears in the training data.
What AI hiring laws and cases should employers know?
iTutorGroup is one item in a fast-growing set of laws, cases, and guidance on AI in employment. The table below summarizes the most documented items.
Item: ADEA enforcement (iTutorGroup settlement)
Jurisdiction: United States (federal).
What it does: First EEOC settlement involving AI hiring discrimination, resulting in a $365,000 settlement.
Status / date: Settled in August 2023.
Item: EEOC AI technical assistance
Jurisdiction: United States (federal).
What it does: Clarifies that Title VII applies to AI-based hiring tools and reinforces that the ADEA and ADA are covered under the EEOC's AI fairness initiative.
Status / date: Issued in May 2023.
Item: NYC Local Law 144
Jurisdiction: New York City.
What it does: Requires independent bias audits of Automated Employment Decision Tools (AEDTs) and advance notice to candidates.
Status / date: Enforcement began in July 2023.
Item: Illinois AI Video Interview Act
Jurisdiction: Illinois.
What it does: Requires candidate notice, an explanation of AI use, and consent before AI analyzes recorded video interviews.
Status / date: Effective January 2020.
Item: Colorado AI Act (SB 24-205)
Jurisdiction: Colorado.
What it does: Establishes consumer protections for high-risk AI systems, including those used in employment decisions.
Status / date: Enacted in 2024; effective date has been delayed and enforcement remains contested.
Item: EU AI Act
Jurisdiction: European Union.
What it does: Classifies AI systems used in hiring and employment as high-risk under Annex III, triggering additional compliance obligations.
Status / date: Entered into force in August 2024, with high-risk requirements phased in over time.
Item: Mobley v. Workday
Jurisdiction: United States (Northern District of California).
What it does: Employment discrimination lawsuit challenging AI-based candidate screening, with the court allowing claims against the vendor under an "agent" theory.
Status / date: ADEA claim conditionally certified as a collective action in 2025.
Two of these items deserve a closer look because they shape day-to-day compliance.
What does NYC Local Law 144 require?
NYC Local Law 144 requires employers and employment agencies using an automated employment decision tool for candidates or employees in New York City to:
Commission an independent bias audit of the tool within the prior year.
Publish a summary of the audit results.
Provide candidates with notice at least 10 business days before the tool is used, including the job qualifications and characteristics the tool assesses.
Enforcement began on July 5, 2023. The law is procedural rather than substantive: it does not ban any tool, but it forces transparency and a documented bias check. An employer who skips the audit faces civil penalties per violation, separate from any underlying discrimination claim.
What is the significance of Mobley v. Workday?
Mobley v. Workday is a federal lawsuit alleging that Workday's AI-based applicant screening discriminated on the basis of age, race, and disability. The plaintiff said he was rejected from a large number of jobs that used Workday's screening, despite being qualified.
In July 2024 the court declined to dismiss Workday and let the claim proceed under an "agent" theory, meaning the software vendor could potentially be held liable as an agent of the employers using its tool. In 2025 the court conditionally certified the ADEA age claim as a nationwide collective action, which lets similarly situated applicants opt in. The case tests whether liability stops at the employer or extends to the vendor.
How can employers prevent AI hiring discrimination?
Preventing algorithmic age discrimination is a governance problem, not only a technical one. The following steps reflect current law and EEOC guidance.
Inventory every tool that touches a hiring decision. This includes resume parsers, ranking algorithms, chatbots, video-interview analyzers, and any third-party AEDT. You cannot govern what you have not catalogued.
Demand the bias audit, do not assume it. Require vendors to produce independent audit results, the data the audit used, and the demographic groups tested. NYC Local Law 144 makes this mandatory for covered employers; treat it as a baseline everywhere.
Test for disparate impact on your own applicant flow. Vendor audits use vendor data. Run the four-fifths rule or a comparable statistical check against your actual applicant pool, broken out by age band, sex, and race.
Remove proxy variables. Graduation year, "digital native," years-of-experience caps, and certain extracurricular markers can stand in for age. Audit feature lists for these proxies.
Keep a human decision-maker accountable. Document who reviews flagged or rejected candidates and on what basis. An auditable human checkpoint reduces the risk of a fully automated rejection that no one can explain.
Preserve records and reasons. Retain application data, model versions, and the basis for adverse decisions. The iTutorGroup reapplication worked because the system left a contradiction in the record; your own records should defend, not indict, your process.
Write notice and consent into the candidate flow. Where Illinois, NYC, or other jurisdictions require it, build notice and consent into the application itself rather than adding it on later.
The cost comparison is direct. A bias audit and a feature review cost a fraction of a settlement, the staff time a consent decree demands, and the reputational damage of a public AI-discrimination case.
Next steps checklist
Action: Catalog all Automated Employment Decision Tools (AEDTs) used throughout the hiring process.
Owner: HR and IT.
Why it matters: Identifies the organization's full scope of AI-related legal and regulatory exposure.
Action: Obtain independent bias audit results from every AI vendor.
Owner: Procurement and Legal.
Why it matters: Supports compliance with NYC Local Law 144 and strengthens vendor due diligence.
Action: Perform a disparate impact analysis on the applicant pipeline.
Owner: People Analytics.
Why it matters: Detects potential age, sex, and race disparities that vendor testing may not identify.
Action: Remove age-related proxy features from screening models.
Owner: Data team or AI vendor.
Why it matters: Reduces the risk of age discrimination claims under the Age Discrimination in Employment Act (ADEA).
Action: Assign a documented human reviewer for all candidate rejections.
Owner: HR.
Why it matters: Ensures meaningful human oversight and avoids unexplained fully automated hiring decisions.
Action: Add candidate notice and consent to the application process where required.
Owner: HR and Legal.
Why it matters: Helps comply with Illinois, New York City, and similar AI employment regulations.
Action: Establish a record retention policy for AI model versions and hiring decisions.
Owner: Legal and IT.
Why it matters: Preserves evidence needed for regulatory compliance and legal defense if a claim is filed.
Frequently asked questions
How much did iTutorGroup pay in the EEOC settlement?
iTutorGroup agreed to pay $365,000 to a group of more than 200 applicants who were rejected because of their age. The 2023 consent decree also required the company to adopt anti-discrimination policies, conduct training, and consider qualified applicants who had previously been screened out. The payment resolved the EEOC's lawsuit without a court ruling on liability.
Was iTutorGroup the first AI hiring discrimination case?
The iTutorGroup matter is widely described as the EEOC's first settlement of an AI-driven hiring discrimination case. Earlier episodes exist, such as Amazon scrapping an internal AI recruiting tool around 2018 after it down-ranked resumes associated with women, but that was an abandoned internal project, not a federal enforcement settlement.
Does the ADEA apply to AI screening tools?
Yes. The ADEA protects workers age 40 and older and contains no exemption for automated tools. If an AI screening system rejects or down-ranks candidates based on age, whether through an explicit cutoff or a biased statistical pattern, the employer using it can be liable. EEOC guidance confirms that existing civil rights laws apply to AI hiring tools.
Who is liable when AI discriminates in hiring, the employer or the vendor?
The employer making the hiring decision is liable, as iTutorGroup's settlement shows; a vendor's "bias-free" claim does not transfer that responsibility. Mobley v. Workday is testing whether liability can also reach the vendor as an agent of the employer. Until that question resolves, employers should assume primary responsibility for any tool they deploy.
