Employment law and AI have collided in just about every part of the employee lifecycle. Hiring tools, performance management, compensation analysis, even termination decisions are being touched by AI. The legal exposure is real, and so is the opportunity.

‍

AI-assisted hiring tools have proliferated, and so have the regulations governing them. New York City's automated employment decision tool law is now several years old; California, Illinois, and Colorado have followed with their own frameworks; the federal government is increasingly active.

‍

Hiring Tools

Companies using AI in hiring need a clear inventory of the tools, the bias audits supporting them, and the disclosures they make to candidates. None of this is optional anymore.

‍

‍

Performance Management

AI in performance management is more contested. Tools that score employee output, monitor productivity, or flag underperformance create exposure under disability discrimination law, privacy law, and — increasingly — collective bargaining agreements.

Our advice to employers is to use these tools as input to human decisions, not as decisions themselves. The legal protection is meaningfully different in those two postures.

‍

Termination Decisions

Termination decisions made on the basis of AI output without human judgment are a high-risk area, full stop. The risk is both substantive (the AI may be wrong or biased in ways that aren't apparent) and procedural (the company's documentation of the decision may not satisfy a court).

We recommend that no termination decision be driven primarily by an AI system, and that the human decision-maker document the reasons in their own words.

‍

‍

Where AI Helps Employers

Used carefully, AI is a powerful tool for employers. It can spot patterns in compensation that suggest pay equity issues. It can flag harassment complaints that share characteristics with prior matters. It can identify retention risks before they become resignations.

Each of these is a positive use — provided the legal and HR teams are involved in the design, not just the cleanup.

‍