New York City has begun enforcing a law that puts hiring algorithms under public scrutiny. Employers using certain automated tools to recruit, hire or promote workers must arrange independent audits and disclose information about the tools and their results.
What employers must disclose
New York City Local Law 144 requires employers using covered algorithms to have them audited by an independent entity. The audit results must be made public, and employees or job candidates must receive disclosures about the tools.
Reports must identify the algorithms in use and show average scores for candidates across different races, ethnicities and genders. A score may take the form of a rating, classification or recommendation.
The reports must also include impact ratios. These compare the average score for people in a particular category with the average score for people in the highest-scoring category. That comparison is intended to make differences in outcomes visible, though the figures alone do not explain why those differences occurred.
The law sets penalties for noncompliance: $375 for a first violation, $1,350 for a second and $1,500 for a third or later violation. Each day an employer uses an algorithm in violation of the law counts separately. Failing to provide sufficient disclosure can also be a violation.
Who the law covers, and what tools may be involved
The law applies when a person is performing or applying for a job in New York City, even if that person does not live in the city. The NYC Department of Consumer and Worker Protection is responsible for enforcement.
Automated hiring tools take many forms. Some scan résumés and cover letters for keywords. Others use chatbots to screen applicants, or analyze speech and facial expressions in an attempt to assess problem-solving skills, aptitudes or “cultural fit.”
These systems can affect people through choices that may seem incidental. The article describes tools that assign different scores based on glasses or a headscarf, penalize Black-sounding names or references to a women’s college, or disadvantage applicants whose disabilities limit keyboard use. The examples show why examining the outcomes of automated screening matters to candidates.
Audits respond to concerns about bias
Recruitment algorithms can carry existing workplace biases into hiring decisions. Khyati Sundaram, CEO of recruitment technology vendor Applied, told TechCrunch that employers should avoid letting AI independently score or rank candidates. She warned that such systems may mirror and perpetuate biases already present in the workplace.
The source points to earlier concerns, including Amazon scrapping a recruiting engine in 2018 after it was found to discriminate against women candidates, and a 2019 academic study showing AI-enabled anti-Black bias in recruiting. It also cites an October 2022 University of Cambridge study suggesting that removing gender and race from assessments may not eliminate bias if ideas about an ideal employee remain shaped by those characteristics.
Meanwhile, AI use in hiring is growing. A February 2022 survey from the Society for Human Resource Management found that nearly one in four organizations used AI to support hiring, with the figure reaching 42% among employers with 5,000 or more employees.
Coverage and impact remain open questions
Critics argue that the law may not reach every relevant tool. The New York Civil Liberties Union (NYCLU) says the law falls short of protecting candidates and workers. In a policy memo, its senior privacy and technology strategist, Daniel Schwarz, raised the possibility that tools such as systems transcribing video or audio interviews could fall outside the law’s scope.
Local Law 144 does require the auditor to be independent: auditors cannot have been involved in using, developing or distributing the algorithm, and cannot have a relationship with the company that submits it. That requirement creates distance between the employer and the audit, while critics continue to call for broader coverage, stronger audit requirements and meaningful notice.
Some employers and technology companies are also exploring voluntary efforts. The Data & Trust Alliance launched in December 2021 to develop an evaluation and scoring system for AI, with a focus on detecting and combating algorithmic bias, particularly in hiring. The source notes that the group included companies such as CVS Health, Deloitte, General Motors, Humana, IBM, Mastercard, Meta, Nike and Walmart at one point.
It is too early to know whether New York City’s law will change hiring outcomes. Its implementation may influence future rules: the source notes proposals under consideration in Washington, D.C., bills introduced in California and a bill introduced in New Jersey. For candidates, the immediate change is that covered tools now come with audit and disclosure obligations; whether those steps are broad and effective enough remains contested.