Using AI in hiring lawfully in the United States
Reviewed by Mellow Editorial Team, HR & payroll content team
Using AI tools to screen resumes, rank candidates, or conduct automated interviews can save time and reduce administrative load. But in the US, those same tools can expose employers to discrimination claims and regulatory scrutiny if deployed without a clear compliance framework.
What the law actually covers
No single federal statute governs AI in hiring specifically. Instead, existing civil rights laws apply. Title VII of the Civil Rights Act, the Americans with Disabilities Act (ADA), and the Age Discrimination in Employment Act (ADEA) all prohibit employment decisions that discriminate based on protected characteristics — regardless of whether a human or an algorithm makes those decisions.
The Equal Employment Opportunity Commission (EEOC) has made clear that an employer cannot escape liability by attributing a biased outcome to a vendor's software. If your AI screening tool systematically deprioritizes applicants of a particular race, gender, or age group, that is your legal problem, not just the vendor's.
The operative legal concept here is disparate impact: a neutral-looking practice that produces disproportionately adverse outcomes for a protected group. AI tools are particularly susceptible to this if they were trained on historical hiring data that already reflected biased human decisions.
State and local laws add layers
Several jurisdictions have moved faster than the federal government. The most prominent example is New York City Local Law 144, which took effect in 2023. It requires employers using automated employment decision tools (AEDTs) to:
- Commission an independent bias audit before using the tool
- Publish the audit results publicly
- Notify candidates that an AEDT is being used and, on request, explain what data it relies on
Illinois enacted the Artificial Intelligence Video Interview Act, which governs AI-analyzed video interviews and requires employer disclosure plus candidate consent.
Maryland and several other states have enacted or proposed similar disclosure requirements. This landscape is moving fast. If you operate across multiple states, track legislation in each one — what is optional in Texas may be mandatory in Illinois.
What AI tools actually do in hiring (and where the risk sits)
"AI in hiring" covers a wide range of tools:
- Resume parsers and rankers that score applicants based on keywords or predicted fit
- Automated video interview platforms that analyze speech patterns, facial expressions, or word choice
- Personality and cognitive assessments with AI-scored results
- Chatbots that conduct initial screening conversations
Risk is not uniform. A tool that simply scans for specific credentials carries different legal exposure than one that infers personality traits from micro-expressions. The further a tool moves from objective, job-related criteria, the harder it becomes to demonstrate that its outputs are valid predictors of job performance — which is the standard courts and regulators will apply.
The EEOC's Uniform Guidelines on Employee Selection Procedures remain the baseline framework: any selection procedure that has adverse impact must be validated as job-related and consistent with business necessity.
Practical steps to reduce legal exposure
You do not need to avoid AI tools entirely. You do need to use them deliberately.
Audit before you deploy. Ask vendors for documentation on how their tool was trained, what its known disparate impact rates are across demographic groups, and whether an independent third-party audit exists. Treat vague answers as a red flag.
Check state requirements first. If you hire in New York City or Illinois, specific legal obligations already apply. Review them before selecting a tool, not after.
Keep humans meaningfully involved. Using AI to generate a ranked shortlist is different from using it to make a final hiring decision. Documenting that a qualified human reviewed AI outputs and made the ultimate call strengthens your position if a decision is challenged.
Preserve records. Retain records of the selection criteria, the tool used, audit results, and individual hiring decisions. If a disparate impact claim arises, you will need this documentation.
Review outcomes periodically. Run your own informal analysis of who your AI-assisted process is and is not advancing. If a pattern emerges — particular groups consistently failing at the same stage — investigate before regulators or plaintiffs do.
The ADA consideration most employers overlook
The ADA introduces a specific wrinkle. Automated tools may disadvantage candidates with certain disabilities — for example, an AI video interview tool may score speech patterns in ways that penalize candidates with speech impairments or processing differences. Employers are generally required to provide reasonable accommodations in the application and interview process. That obligation extends to AI-assisted steps: if a candidate requests an accommodation to participate in a different format, you need a process to handle that request.
This is worth building into your intake workflow explicitly, not leaving as an afterthought.
Staying current
Federal AI regulation specific to employment remains in development. The EEOC has issued technical guidance and signaled ongoing attention to algorithmic discrimination. State legislatures are active. Employers using AI-assisted hiring across multiple jurisdictions should designate someone to track regulatory changes — or engage employment counsel to do so — because the rules in effect today may look different in twelve months.
---
Run HR and payroll in United States with Mellow
Mellow brings HR, payroll and 12 AI agents into one platform — built to handle United States properly, with payroll included, from £4 per employee per month. The AI agents don't just answer questions; they generate contracts, run cost estimates and draft letters for you.
- United States payroll software
[Start a free trial →](/register)