Updated May 23, 2026
Disparate Impact in Hiring: Why Your Cognitive Tests May Violate California Law
Disparate impact occurs when seemingly neutral hiring practices disproportionately exclude protected groups, and cognitive tests are increasingly triggering legal violations under California law. Many employers rely on pre-employment assessments, AI screening tools, and personality evaluations without realizing these tests may create unintended discrimination. California's new automated decision system regulations, in fact, impose strict requirements on businesses using cognitive testing in hiring. This article explains which cognitive tests are covered under California law, how disparate impact applies to your hiring practices, and the specific compliance steps you must take to avoid legal liability.
What Cognitive Tests Are Covered Under California Law
California law applies to a broader range of testing tools than most employers realize. The regulations cover traditional paper-and-pencil assessments alongside modern algorithmic screening systems, creating compliance obligations across your entire hiring technology stack.
Pre-employment cognitive assessments
Aptitude tests, sometimes called cognitive ability tests, examine problem-solving, verbal, mathematical, or decision-making skills. These assessments correlate well with job performance but measure ability in isolation. They can't assess an applicant's other attributes, such as personality or collaboration style.
Cognitive tests evaluate aspects of intelligence without being traditional IQ tests. Employers use them to predict how well a candidate will perform in a role. The tests measure reasoning, memory, and problem-solving capabilities. Common questions assess critical thinking, learning abilities, stress tolerance, and verbal communication skills.
While not obviously discriminatory, cognitive tests can illegally impact groups of workers and job applicants by discriminating against particular classes. When a test produces a disproportionate negative impact on a particular group, employers must explore alternative evaluation methods.
AI-powered screening tools
Automated decision systems represent the newest category of cognitive testing under California regulations. These tools use algorithms to evaluate candidates based on data inputs and predetermined criteria.
AI-powered screening encompasses video interview analysis platforms, resume parsing systems, and predictive algorithms that score applicant suitability. The technology may analyze speech patterns, word choices, or behavioral indicators captured during digital interactions. These systems fall squarely within California's regulatory framework for automated employment decision systems.
Personality and psychological evaluations
Personality tests claim to measure behavioral factors and traits. Employers use these assessments to assemble complementary teams with matching personality characteristics. In reality, personality assessments don't come from hard science.
The Big Five personality traits—Conscientiousness, Agreeableness, and Emotional Stability—are measured most frequently in employment contexts. These traits predict various work behaviors and job performance indices across occupations. A 2003 survey indicated that about 30% of all companies use personality tests to assist in employment decisions.
Personality tests face several validity concerns. Some have correct responses that are easy to guess from question tone. Well-known assessments like the Myers-Briggs Type Indicator can give different results on different days. Most tests use a yes/no format poorly suited to personality questions where the answer is often "sometimes" or "most of the time".
Pre-employment evaluations consist of record reviews, clinical interviews, and personality testing. Psychological testing measures traits such as honesty, dependability, and emotional stability. A personality test can lawfully be administered prior to a conditional offer if designed to assess normal-range traits exclusively, rather than identify mental or emotional impairment.
Skills and aptitude tests
Job skills tests, also called knowledge tests, analyze what the applicant knows about the job and their skill level. Two distinct categories exist:
Hard skills are technical competencies requiring training or certification. Coding, copywriting, and marketing fall into this category.
Soft skills are non-technical competencies like management ability, leadership, and attention to detail. These skills may be desirable or essential depending on the position.
Skills testing shows what candidates already know, not how quickly they can learn or how they fit company culture. The testing is less important for entry-level positions where employees learn on the job.
Understanding Disparate Impact in Hiring
The legal framework separating lawful and unlawful hiring practices hinges on understanding two distinct forms of discrimination. Disparate treatment involves intentional discrimination against individuals based on protected characteristics. Disparate impact, conversely, refers to unintentional discriminatory practices where seemingly neutral policies cause disproportionate harm to protected groups.
What disparate impact means
Disparate impact discrimination occurs when facially neutral policies or actions produce unjustified negative effects on specific groups, regardless of employer intent. The concept gained legal recognition in 1971 when the Supreme Court decided Griggs v. Duke Power Co., establishing that Title VII prohibits not only overt discrimination but also practices that appear fair yet operate discriminatorily.
In Griggs, a power company required high school diplomas for certain departments, a policy that disproportionately eliminated Black applicants despite lacking significant relation to job performance. The Court held this violated Title VII because the requirement wasn't related to successful job performance and didn't advance the company's stated goals.
Equal Employment Opportunity Commission regulations use the 80% rule to identify potential disparate impact. Under this standard, selection rates for protected groups must reach at least 80% of the selection rate for the most selected group. When a height requirement excludes far more women than men, as an illustration, the employer must demonstrate the requirement serves a substantial, legitimate business need.
The burden-shifting framework proceeds in three steps. Plaintiffs first demonstrate that a specific employment practice disproportionately impacts a protected group through statistical analysis. The burden then shifts to employers to prove the practice is job-related and consistent with business necessity. Even when employers establish business necessity, plaintiffs can still prevail by showing less discriminatory alternatives exist that meet the employer's needs.
How cognitive tests create unintended discrimination
Cognitive ability assessments demonstrate how neutral practices create discriminatory effects. These tests, administered identically to all candidates, can disproportionately screen out members of protected groups while appearing objective. Autistic and neurodivergent job applicants face significant employment barriers from cognitive assessments that fail to accommodate diverse needs.
Tests measuring memory and cognitive processing show racial disparities in performance. Black job candidates and other candidates of color face disadvantages from cognitive ability assessments that have long demonstrated discriminatory screening patterns. The technology compounds existing problems when combined with AI tools, exacerbating harmful employment barriers based on race, gender, disability, and other protected characteristics.
Protected characteristics under California law
California's Fair Employment and Housing Act protects individuals from discrimination based on race, color, ancestry, national origin, religion, age (40 and over), disability (mental and physical), sex, gender (including pregnancy, childbirth, breastfeeding), sexual orientation, gender identity, gender expression, medical condition, genetic information, marital status, military or veteran status, and reproductive health decision-making. These protections apply to employers with five or more employees, covering all aspects of employment from hiring through termination.
California's Automated Decision System Regulations
California finalized regulations addressing automated decision systems in employment, with provisions taking effect October 1, 2025. A separate set of requirements becomes effective January 1, 2027, imposing documented risk assessments and pre-use notices for many employers. These rules apply to all employers covered by the Fair Employment and Housing Act, meaning those with five or more employees.
Key definitions employers must know
An automated decision system means any computational process that makes a decision or facilitates human decision making regarding an employment benefit. The definition encompasses artificial intelligence, machine learning, algorithms, statistics, and other data processing techniques. This covers systems used for hiring, allocation of work, compensation, promotion, demotion, suspension, and termination.
The regulations define AI as a machine-based system that infers from input how to generate outputs, including predictions, content, recommendations, or decisions. A proxy is any characteristic or category closely correlated with a protected category under FEHA. Medical or psychological examination includes any test, question, puzzle, game, or other challenge likely to elicit information about a disability, even when administered through an automated decision system.
Human review requirements specify that reviewers must know how to interpret the system's outputs, analyze those outputs alongside other relevant information, and possess authority to make or change the decision based on that analysis. In practice, this requires both policies ensuring adequate human review and training for reviewers.
Prohibited practices under the new rules
Employers cannot use automated decision systems or selection criteria that discriminate against applicants or employees based on protected characteristics. Facially neutral technology becomes unlawful if it produces adverse impact on protected groups, unless the employer demonstrates the practice is job-related and consistent with business necessity.
The rules specify that preventing discrimination includes offering reasonable accommodation for disability or religious needs connected with system use. Employers must preserve personnel and employment records for four years instead of two. This requirement extends to automated decision system data, defined as any data used in or resulting from a system, plus data used to develop or customize a system for employer use.
When cognitive tests are presumed unlawful
Systems measuring reaction time, skill, dexterity, or other abilities may discriminate against individuals with certain disabilities or protected characteristics. Tools analyzing tone of voice, facial expressions, or physical characteristics may discriminate based on race, national origin, gender, disability, or other protected traits. Schedule availability screening can disproportionately impact individuals with religious obligations, disabilities, or medical conditions.
Business necessity defense requirements
Employers defending disparate impact claims must prove the practice serves a purpose to operate the business safely and efficiently, and that the practice substantially accomplishes this business purpose. The test requires an overriding legitimate business purpose making the practice necessary to safe and efficient operations. The challenged practice must effectively carry out the stated business purpose, and no acceptable alternative policies or practices can better accomplish the purpose, or accomplish it equally well with lesser differential impact.
Evidence of anti-bias testing or similar proactive efforts to avoid unlawful discrimination is relevant to any claim or defense. Courts and agencies may consider the quality, scope, recency, results, and employer response to bias testing. The absence of such evidence may weigh against employers choosing not to evaluate their systems.
Common Cognitive Tests That Violate California Law
Automated tools marketed as cost-effective hiring solutions often violate California law by creating disparate impact on protected groups. While vendors promote these systems as objective and efficient, the technology frequently replicates historical biases and screens out qualified candidates.
Tests measuring reaction time and physical responses
Systems measuring skill, dexterity, reaction time, or other physical abilities may discriminate against individuals with certain disabilities. These assessments disadvantage candidates with mobility limitations, neurological conditions, or speech differences. California regulations require employers to provide reasonable accommodations during the assessment process, allowing applicants with disabilities to be evaluated accurately and fairly. Without accommodations, qualified candidates who could perform the job successfully get screened out before human review occurs.
Video interview analysis tools
Platforms evaluating candidates through facial expressions, speech patterns, or vocal tone present discrimination risks across multiple protected categories. Tools analyzing an applicant's tone of voice, facial expressions, or other physical characteristics may discriminate based on race, national origin, gender, or disability. Video interviewing systems that incorporate AI elements are likely to discriminate based on disability, race, and other protected characteristics. The automated features exacerbate fundamental problems by assessing behaviors through a narrow lens that fails to account for neurological differences or cultural communication styles.
Schedule availability screening
Tools ranking or screening applicants based on schedule availability disproportionately impact individuals with religious obligations, disabilities, or medical conditions. These systems exclude candidates who observe religious sabbaths or have caregiving obligations. Schedule-based screening causes disparate impact despite appearing neutral on its face.
Personality prediction algorithms
Products claiming to predict employee performance or behavior based on personality, beliefs, or emotional states pose significant legal risks. Assessments measuring general personality traits such as positivity, emotional awareness, liveliness, ambition, and drive are not clearly job related or necessary for specific jobs. These tools unfairly screen out people based on disabilities. Individuals with post-traumatic stress disorder may perform poorly on personality assessments despite being able to perform the job successfully.
Accent and language-based assessments
Employment discrimination based on an applicant's accent is unlawful unless the employer proves the individual's accent materially interferes with the ability to perform the job. Similarly, discrimination based on English proficiency is unlawful unless justified by business necessity. The regulations prohibit discrimination tied to accents unless they materially interfere with job performance.
Compliance Steps for Employers Using Cognitive Tests
Employers using cognitive tests must implement specific compliance measures to avoid disparate impact liability under California regulations. These requirements apply whether you develop assessment tools internally or purchase them from third-party vendors.
Conducting anti-bias testing
The regulations identify six factors relevant to anti-bias testing or similar proactive efforts to avoid unlawful discrimination. Employers should evaluate the quality of testing efforts, the efficacy of the tests, and the recency of testing. Equally important are the scope of tests, what results reveal about system outcomes, and whether and how you responded to the results and their implications. A single validation at launch does not demonstrate sufficient reasonable measures.
Recordkeeping requirements for four years
Employers must preserve personnel and employment records for four years instead of two. This requirement extends to automated decision system data, including any data used in or resulting from system application, data provided by or about individual applicants or employees, data reflecting employment decisions or outcomes, and any data used to develop or customize a system for your use.
Reviewing third-party vendor agreements
Employers cannot outsource accountability. Conduct robust due diligence on vendor tools to determine whether you can use them in compliance with regulations. Your review should evaluate the tool's design, inputs, outputs, and validation processes, including any evidence of bias testing or disparate impact analysis.
Providing reasonable accommodations
You must provide reasonable accommodation to applicants or employees consistent with disability and religious creed protections. Accommodations ensure individuals with disabilities can demonstrate their true aptitude rather than their disability.
Training human decision-makers
Human reviewers must know how to interpret system outputs, analyze those outputs alongside other relevant information, and possess authority to make or change decisions based on that analysis. Design decision flows that invite human review, challenge, correction, and intervention.
Conclusion
Cognitive testing in hiring carries significant legal risks under California's evolving regulations. By and large, employers underestimate how seemingly neutral assessments create disparate impact on protected groups. The regulatory framework now requires documented anti-bias testing, four-year recordkeeping, reasonable accommodations, and meaningful human review of automated decisions.
Vendor-provided tools don't absolve you of liability. Indeed, your organization remains accountable for discriminatory outcomes regardless of who developed the assessment. Start by auditing your current testing practices against the prohibited categories outlined here. Then implement the compliance steps before the 2027 requirements take effect. A proactive approach protects both your organization and the qualified candidates your tests might otherwise exclude unfairly.
References
https://www.findlaw.com/employment/hiring-process/tests-at-work-your-rights.html
https://calcivilrights.ca.gov/wp-content/uploads/sites/32/2018/05/FinalTextRegNationalOriginDiscrimination.pdf
https://www.disabilityrightsca.org/publications/discrimination-fact-sheet-testing-and-licensing-accommodations
Call Setyan Law at (213)-618-3655 for a consultation.
