What New York employers must know about disparate impact discrimination
Governor Hochul signed a law in December 2025 that holds employers liable for neutral policies that disproportionately harm protected groups, even without discriminatory intent.
On December 19, 2025, Governor Kathy Hochul signed Senate Bill S8338, amending the New York State Human Rights Law to expressly allow discrimination claims based on disparate impact. The law applies to employment discrimination occurring on or after that date. While New York courts had long treated disparate impact as actionable, the amendment codifies the standard into statute, establishing formal rules that employers must now follow.
Like federal law under Title VII, this amendment establishes liability based on disparate impact. A neutral hiring policy, background check, or compensation structure can now trigger liability if it disproportionately harms workers based on race, gender, disability, age, religion or other protected characteristics—even if the employer had no discriminatory intent.
How a facially neutral policy creates liability
The new law holds that 'a facially neutral employment practice may violate the New York State Human Rights Law based on its discriminatory effects, even absent discriminatory intent.' This means an employer can face liability for a policy that appears neutral on its face but produces unequal outcomes for a protected group.
For example, a background check policy that bars all applicants with criminal histories may have a disparate impact on certain populations, even though the policy applies equally to everyone. Another example: a scheduling system that rotates shifts without regard to caregiving obligations may disproportionately harm workers with disabilities who need consistent schedules for medical appointments.
The law requires no proof that an employer intended to discriminate. An employee only needs to show that a specific policy causes—or will foreseeably cause—a disparate effect on a protected group. The inclusion of 'predictably' alongside 'actually' means employers must also consider whether a policy is likely to produce unequal results, not just whether it already has.
The three-step burden-shifting framework
Once an employee demonstrates disparate impact, the law creates a formal three-step framework that allocates burdens between the parties. The employee starts: they must show that a specific policy creates a discriminatory effect on a protected group. Statistical evidence matters here, but a mere imbalance between protected and nonprotected workers is not enough without evidence that an identifiable practice caused it.
If the employee meets this burden, the employer then bears responsibility to prove that the policy is 'job-related for the position in question and consistent with business necessity.' Importantly, the law specifies that this justification 'be supported by evidence and not based on hypothetical or speculative considerations.' An employer cannot claim a practice is necessary for business without concrete documentation.
Even if an employer proves business necessity, the employee gets a final opportunity to prevail by showing the employer's goal 'could be satisfied by a less discriminatory alternative.' If a different policy would serve the employer's legitimate business need while producing fewer disparate effects, the original policy is unlawful. This means employers cannot defend policies simply by asserting they are necessary; they must also show no better option exists.
Hiring practices and algorithmic tools face heightened scrutiny
The amendment carries particular consequences for hiring decisions and the automated systems employers use to make them. Algorithmic screening tools, AI-powered resume reviewers, and machine-learning systems used in hiring can all trigger disparate impact liability if they produce unequal outcomes.
An AI system might, for instance, be trained on historical hiring data that reflects past discrimination. If the system learns to favor applicants with names that were common among previously hired groups, or educational backgrounds more accessible to certain demographics, the system can create disparate impact even if the employer never programmed in discrimination. The employer would then need to prove the AI tool is job-related and meets business necessity—a difficult standard when the tool is not fully transparent even to the employer.
Employers using algorithmic hiring tools should carefully document why they chose the system and what results it produces across demographic groups. Proactive audits to detect disparate outcomes before they become the basis for legal claims are now essential risk management. Some employers may find it prudent to run parallel hiring processes—one using the automated tool, one using traditional methods—to measure whether the tool produces different outcomes.
“An employer cannot claim a practice is necessary for business without concrete documentation.”
What the law requires employers to review
Employment practices that warrant close examination include background check policies; compensation and bonus structures; shift scheduling systems; leave and accommodation policies; performance evaluation criteria; and hiring or promotion requirements. A policy that appears neutral—such as requiring a college degree for a job that does not truly require one—can be challenged if fewer members of a protected group hold the required credential.
Disability accommodation presents a particular area of risk. For instance, policies that fail to provide necessary accommodations for workers with disabilities can have disparate impact. This includes restrictions on service animals without accommodation exceptions, leave caps without medical exceptions, and inflexible scheduling systems.
The law does not prohibit these practices entirely. Rather, employers can defend them by showing they are job-related and necessary, and that no less discriminatory alternative exists. A restaurant might defend a 'no animals' policy (service dogs excluded) as necessary for health code compliance. An employer might defend maximum leave caps as necessary to ensure adequate staffing. The key is documenting the business reason and considering whether the policy could be redesigned to achieve the same goal with less disparate effect.
How this compares to federal law and prior New York practice
The new New York law codifies what federal Title VII law already allows under disparate impact theory. Both laws permit claims based on effect rather than intent. However, the New York City Human Rights Law, which predated the state amendment, already had an even broader disparate impact standard that applied to employment. The new state law brings the broader state law into alignment.
One notable feature: the NYCHRL had created a lower threshold than federal law, requiring employers to show that a policy 'bears a significant relationship to a significant business objective' rather than meeting the stricter 'business necessity' standard of federal law.
