Unveiling NYC's Algorithm Hiring Law: Understanding Bias Audits

Unveiling NYC's Algorithm Hiring Law: Understanding Bias Audits

Table of Contents

1. Introduction
2. Understanding New York City Local Law 144
3. Scope and Obligations of the Law
4. Bias Audits: Definition and Requirements
5. Independent Auditors: Roles and Responsibilities
6. Clarifications on Automated Employment Decision Tools (AEDTs)
7. Implications for Vendors and Third-Party Providers
8. Compliance Challenges and Considerations
9. Enforcement of the Law
10. Conclusion

👉Introduction

In this article, we will delve into the intricacies of New York City Local Law 144, focusing on the requirements, implications, and challenges surrounding algorithmic auditing and bias audits for automated employment decision tools (AEDTs). This law mandates that employers using AEDTs must conduct unbiased audits to mitigate disparate impact and ensure fair decision-making criteria for job applicants. With January 1st, the deadline for compliance, looming, it is crucial for companies to understand the law's provisions, navigate the complexities, and prepare for enforcement. Let's dive in and explore the details.

👉Understanding New York City Local Law 144

New York City Local Law 144 was enacted to address the potential bias and discriminatory outcomes arising from the use of AEDTs in employment decision-making processes. It places several obligations on employers and employment agencies utilizing AEDTs to ensure fairness and reduce adverse impact. The law outlines the duty of employers to conduct regular bias audits, provide public summaries of the audits, and notify candidates about the use of AEDTs in the hiring process.

👉Scope and Obligations of the Law

Under the law, employers using AEDTs within the New York City jurisdiction or assessing candidates residing in New York City are subject to its obligations. The law emphasizes the responsibility of employers and employment agencies, as they are ultimately accountable for adhering to the law's requirements. The core obligations include conducting independent bias audits, calculating selection rates and impact ratios for various categories, and making the bias audit results public.

👉Bias Audits: Definition and Requirements

A bias audit is an impartial evaluation conducted by an independent auditor to assess the disparate impact and adverse impact of an AEDT's usage. It focuses on identifying bias and unfairness in the employment decision-making process, particularly related to race, gender, and other legally protected categories. Bias audits must be performed annually and include the calculation of selection rates and impact ratios for each category under review.

👉Independent Auditors: Roles and Responsibilities

To ensure the integrity and impartiality of bias audits, the law defines an independent auditor as a person or group that is not involved in using or developing the AEDT. The independence of the auditor is crucial to maintain objectivity and prevent conflicts of interest. Employers have the responsibility to engage independent auditors to conduct the bias audits and ensure complete transparency in the process.

👉Clarifications on Automated Employment Decision Tools (AEDTs)

The law provides Clarity on the definition of AEDTs, emphasizing their reliance on machine learning, statistical modeling, data analytics, and artificial intelligence. It further refines the concept of "substantially assist or replace discretionary decision-making," clarifying that AEDTs either solely rely on Simplified outputs for decision-making or use such outputs as the primary criterion in an automated decision-making system. The law also addresses situations where AEDTs modify or overrule human decisions.

👉Implications for Vendors and Third-Party Providers

As employers often utilize AEDTs developed by third-party vendors, the law raises questions about the responsibilities and obligations of these vendors. Though not explicitly stated, the law implies that vendors can assist in conducting bias audits upon employer request. However, the issue of impartiality and independence must be carefully considered. Vendors should be transparent about the functioning of their AEDTs, ensuring compliance with the law's requirements.

👉Compliance Challenges and Considerations

Complying with New York City Local Law 144 poses several challenges for employers and vendors alike. The short timeline until the law's enforcement requires Prompt action to assess AEDTs, conduct bias audits, and implement necessary changes. Employers must grapple with issues of data privacy, statistical significance, and intersectional biases. Adequate resources, robust processes, and effective communication between employers and vendors are vital for successful compliance.

👉Enforcement of the Law

New York City's Department of Consumer and Worker Protection serves as the enforcement agency for Local Law 144. The extent of their resources and enforcement strategy remains unclear. However, the law's requirement for public summaries of bias audits enhances transparency, making it easier for authorities to monitor compliance. Non-compliance can lead to penalties, underscoring the importance of diligently adhering to the law's provisions.

👉Conclusion

New York City Local Law 144 emphasizes the significance of algorithmic auditing, bias audits, and responsible AI in employment decision-making. Employers must actively address the potential biases in their AEDTs, conduct regular bias audits, and ensure compliance with the law. Transparency, impartiality, and effective collaboration between employers and vendors are crucial for successfully navigating the complexities of the law. By understanding and adhering to the provisions of Local Law 144, companies can foster fair and unbiased employment practices in the era of AI.

Resources:

• New York City Local Law 144