AI Hiring: So, You Want C-3PO to Vet Your Candidates?

It may be tempting to let your protocol droid vet the résumé pile, but AI-powered hiring tools can create real-world risks that are anything but science fiction. Employers increasingly use AI to screen resumes, rank candidates, and analyze video interviews to identify perceived “best fit” applicants. These tools can offer hyperspace-level speed to screen candidates, but they can also create a false sense of objectivity. The underlying models may incorporate biases that are not readily apparent and may not be disclosed by the vendor.

Regulators are paying close attention to the use of AI in hiring and its potential impact on equal employment opportunity. Employers should not assume their “Protocol Droid” gets a free pass just because it is powered by sophisticated large language models or other AI systems designed for recruiting. For example, the U.S. Equal Employment Opportunity Commission (EEOC) has emphasized that longstanding anti-discrimination laws apply to algorithmic tools used in employment decisions, including tools designed or administered by third-party vendors (see EEOC, Select Issues on Adverse Impact in AI Employment Selection (2023); 29 C.F.R. pt. 1607). New York City requires certain automated employment decision tools to undergo bias audits and for employers to provide notice to candidates before using these tools (N.Y.C. Admin. Code § 20-870 et seq.). Illinois law imposes disclosure, explanation, consent, and deletion requirements for AI-analyzed video interviews (820 ILCS 42/1 et seq.).

Employers should carefully evaluate an AI hiring tool’s underlying training data, development methodology, and testing processes to understand the potential for biased or discriminatory outcomes. AI systems trained on historical hiring data may learn and perpetuate existing patterns of bias. Likewise, training datasets that are incomplete or lack sufficient diverse representation may produce outcomes that disproportionately exclude individuals with protected characteristics. Such outcomes can cause qualified candidates to be overlooked and expose employers to legal, regulatory, and reputational risks.

The bottom line: AI hiring tools should assist, not replace, human judgment. Employers considering these tools should understand how they work, evaluate them for disparate impact, provide any required notices to candidates, and maintain meaningful human oversight. Used responsibly, AI hiring tools can help streamline recruiting.   Used blindly, they can turn a streamlined hiring process into a compliance, reputational, and fairness problem faster than a jump to lightspeed.

Key takeaway: AI can be a helpful tool in the recruitment and hiring process, but employers remain responsible for ensuring employment decisions are fair, transparent, and compliant with applicable anti-discrimination laws.