In the United States, the use of AI in recruitment does not remove employers’ responsibilities under employment discrimination law. Existing legal protections can apply when automated tools influence hiring decisions.
Title VII of the Civil Rights Act of 1964
Title VII prohibits covered employers from discriminating (U.S. Equal Employment Opportunity Commission [EEOC], n.d.) because of race, color, religion, sex, or national origin. These protections apply to covered employment decisions, including hiring. An employer does not automatically avoid its legal obligations by using an algorithm or purchasing software from a vendor.
Understanding disparate impact
A seemingly neutral employment practice may raise disparate-impact concerns when it disproportionately disadvantages people in a protected group and cannot be justified under the applicable legal standards. An AI screening tool could raise concerns if it systematically excludes qualified applicants from a protected group.
A disparity alone does not automatically establish unlawful discrimination. The legal analysis depends on the evidence, the selection procedure, the applicable standards, and the employer’s justification. Nevertheless, employers should not assume that software-generated decisions are lawful simply because a person did not make each decision directly.
EEOC guidance and other considerations
The The U.S. Equal Employment Opportunity Commission (EEOC) issued technical assistance in 2023 (EEOC, 2023) addressing adverse impact in software, algorithms, and AI used in employment selection procedures. This guidance explains how established Title VII principles apply to automated systems; it is guidance on existing law, not a separate statute governing every AI hiring tool.
The Americans with Disabilities Act may also be relevant when employment assessments disadvantage applicants with disabilities or when reasonable accommodations are required. Privacy and data-protection obligations may apply depending on the information collected, the tool, and the jurisdiction. State and local laws may impose additional requirements.
Are current laws sufficient?
Existing laws provide an important foundation, but enforcement can be challenging when applicants do not know that an algorithm influenced a decision, vendors control technical information, or employers fail to test outcomes. Legal compliance also does not necessarily address every ethical concern about transparency, dignity, or public trust.
My legal conclusion: Existing protections are essential, but employers must implement them effectively through appropriate testing, documentation, investigation of potential adverse impact, and clear accountability. The NIST AI Risk Management Framework offers voluntary guidance (Tabassi, 2023); it is not itself an employment discrimination statute.
How protections relate to AI-assisted hiring
These sources have different legal status: statutes create legal obligations; NIST’s framework is voluntary guidance.
LEGAL AND GOVERNANCE MAP
Different protections serve different roles
| Authority or framework | What it contributes | What it does not automatically do |
|---|---|---|
| Title VII of the Civil Rights Act | Prohibits employment discrimination based on sex and other protected characteristics; disparate-impact principles may be relevant | Does not certify a hiring algorithm as fair before use |
| EEOC technical assistance | Explains how existing employment-discrimination rules can apply to algorithmic tools | Is not itself a new statute or a blanket approval of a vendor’s tool |
| NIST AI Risk Management Framework | Offers voluntary practices for identifying, assessing, and managing AI risks | Does not replace legal obligations or guarantee compliance |
General educational summary. Applicability depends on the facts, the tool, and the relevant law; consult the primary sources listed on the References page.
Sources: U.S. Equal Employment Opportunity Commission (n.d., 2023); Tabassi (2023). See the full references.