New Jersey’s Guidance on AI Use in Hiring
On January 9, 2025, New Jersey’s Attorney General and the Division on Civil Rights (DCR) introduced a significant initiative aimed at addressing the implications of artificial intelligence (AI) in hiring practices. This initiative, known as the Civil Rights and Technology Initiative, seeks to mitigate the risks of discrimination and bias-based harassment that may arise from the use of advanced technologies.
Overview
The DCR issued guidance regarding the application of the New Jersey Law Against Discrimination (LAD) concerning the discriminatory implications of AI. This guidance is particularly crucial for employers utilizing AI for screening applicants and assessing employee performance.
Algorithmic Discrimination
The guidance clarifies that New Jersey’s LAD applies to algorithmic discrimination. This term refers to discrimination that results from an employer’s use of AI or automated decision-making tools. Notably, employers can be held liable under the LAD even if they did not develop the AI tool or are unaware of its discriminatory effects.
Employers must therefore exercise caution in their use of AI to avoid potential liabilities associated with algorithmic discrimination.
Disparate Treatment and Disparate Impact Discrimination
The guidance outlines examples of algorithmic discrimination, emphasizing that AI tools can facilitate disparate treatment discrimination if they are designed to treat members of a protected class differently. For instance, if an employer selectively uses AI to evaluate only Black applicants while excluding others, this constitutes discriminatory practice. Moreover, even if an AI tool does not directly consider protected characteristics, it may still violate discrimination laws by using proxies closely related to such characteristics, like race or sex.
Furthermore, AI tools may also result in disparate impact discrimination, which occurs when ostensibly neutral criteria unintentionally have a disproportionately negative effect on a protected class. An example highlighted in the guidance is a company that employs AI to assess contract bids, which may inadvertently screen out bids from women-owned businesses.
Reasonable Accommodations
The guidance warns that the use of AI may violate the LAD if it obstructs the provision of reasonable accommodations. For instance, AI systems used in hiring may disproportionately exclude applicants who could fulfill job requirements with reasonable adjustments. Similarly, if AI is utilized to monitor employee productivity, it may unfairly flag employees who receive additional break time due to disabilities.
Liability
A critical aspect of the guidance is its broad interpretation of liability concerning algorithmic discrimination. Under the LAD, employers cannot shift liability to AI vendors or external developers, even if they lack comprehensive knowledge of the AI tool’s functioning.
Best Practices
To mitigate the risk of liability under the LAD, the guidance recommends that employers implement several best practices:
- Establish an AI oversight group that includes representatives from various departments, such as legal, human resources, and IT.
- Develop AI-related policies and procedures.
- Provide training on AI tools and algorithmic bias, restricting usage to trained personnel.
- Conduct thorough vetting of AI vendors and tools.
- Secure appropriate contract provisions from AI vendors, ensuring compliance with applicable laws and understanding of the algorithm’s operations.
- Promptly address issues identified during audits or tests.
- Review employment practices liability insurance for applicable coverage.
- Maintain a human element in decision-making processes involving AI tools.
The guidance underscores the importance of ethical practices in AI utilization within employment contexts, encouraging employers to be proactive in addressing potential biases and ensuring compliance with discrimination laws.
