Regulating the Robots: NYC Mandates Bias Audits for Ai-Driven Employment Decisions

Abstract

The mandating, by New York City, of bias audits for AI-driven employment decisions represents a significant step-change in the regulatory agenda towards greater identification and mitigation of bias in the ethically critical field of recruitment, representing a case study for broader global debates on regulating AI. In this brief, we provide an exposition of the regulation, serving two agendas: firstly providing a summary for those unfamiliar with the legislation aimed at ensuring that AI-driven recruitment tools do not prevent the hiring of a diverse workforce; and secondly, providing a commentary that discusses some shortcomings with the legislation and what it might signal for the future. Our audience is those seeking an introduction and wishing to forward the discussion via critical evaluation. Our primary takeaways are: i. the need for examples; ii. clarification about what ‘compliant’ would look like; and iii. the need for a more standardised approach.

Keywords: Bias, bias mitigation, transparency, artificial intelligence, algorithm, recruitment, psychometrics