A PYMNTS Company

NYC Companies Shirk Algorithm Transparency Mandate in Hiring

 |  January 23, 2024

A recent report by The Wall Street Journal highlights that a significant number of companies in the city are overlooking a six-month-old law requiring them to disclose the impact of algorithms on hiring and promotion decisions. According to a study by Cornell University, only 18 out of 400 companies utilizing “automated employment decision tools” have published results on their websites as mandated by Local Law 144.

Local Law 144, enacted to address concerns of potential bias in technology-driven employment decisions, was designed to encourage companies to assess whether their algorithms unintentionally exhibit bias. However, the low compliance rate suggests that the law is not having the intended impact on the majority of employers.

The Cornell University study points out that the wording of the law may be a contributing factor to the lack of adherence. One researcher, who participated in the study, argued that the law provides employers with “almost unlimited discretion” when deciding whether they are complying with the regulation or not.

Advocates of algorithmic transparency in employment processes argue that it is crucial to identify and rectify any biases that may exist in automated decision-making systems. These systems often rely on historical data, which can perpetuate existing inequalities and result in biased outcomes.

The low compliance rate raises questions about the effectiveness of Local Law 144 and whether it provides sufficient guidance for companies to meet its objectives. Critics argue that without clearer guidelines and more stringent enforcement, the law may not succeed in its mission to promote fairness and accountability in algorithmic hiring and promotion practices.

As the debate surrounding the impact of algorithms on employment decisions continues, the city may face pressure to revisit and potentially revise the legislation to ensure companies take proactive measures to address any biases in their automated systems.

Source: Linkedin