Restricts the use by an employer or an employment agency of electronic monitoring or an automated employment decision tool to screen a candidate or employee for an employment decision unless such tool has been the subject of an impact assessment within the last year; requires notice to employment candidates of the use of such tools; provides remedies for violations.
S10147 would add a new article to the Labor Law, titled the “Boundaries on Technology Act,” to regulate employer use of automated employment decision tools in hiring and other workplace decisions. The bill applies to employers with 100 or more employees and generally prohibits use of such tools unless they have undergone a recent impact assessment by an impartial auditor. Those assessments must identify the tool’s inputs, modeling techniques, and any potential disparate impact on protected classes, and must be repeated annually while the tool remains in use.
The bill also requires employers to give clear notice to employees and job candidates when an automated tool is used, including information about what data is collected, what the tool evaluates, and the results of the most recent impact assessment where applicable. It limits employers from relying solely on automated outputs for decisions such as hiring, promotion, discipline, termination, or compensation, and requires meaningful human oversight. The bill further creates rights for current and former employees to access and correct data used by these tools, and it prohibits retaliation against workers who request information or file complaints.
In addition to regulating employers, the bill imposes notice obligations on vendors that sell or distribute automated employment decision tools, requiring them to inform employers of the law’s requirements and provide information about bias or disparate-impact assessments. Enforcement authority would rest with the Labor Commissioner and the Attorney General, with civil penalties and other remedies available for violations. The bill would also amend Labor Law section 218 to add violations of the new article to the commissioner’s enforcement and penalty framework.
The overall sentiment reflected in the bill text and available context is protective of workers and candidates, with a strong emphasis on transparency, anti-discrimination safeguards, privacy, and human review. Because there are no committee transcripts or recorded votes in the provided materials, there is no documented opposition or support from debate history; the bill appears to be introduced as a consumer- and worker-protection measure aimed at limiting unchecked algorithmic decision-making in employment.
The main points of contention likely concern compliance costs, the scope of required audits and disclosures, the burden on employers and vendors, and how to define or measure disparate impact in automated systems. Another likely issue is the bill’s broad treatment of employee data, including biometric, health, communications, geolocation, and workplace-performance information, which may raise privacy and operational concerns for employers while being viewed by supporters as necessary to detect bias and ensure accountability.
The bill would create a new Labor Law article 35-A regulating automated employment decision tools and related electronic monitoring practices in employment decisions. It would impose pre-use and annual impact-assessment requirements, notice obligations, data-access and correction rights, anti-retaliation protections, and vendor disclosure duties, while also authorizing enforcement by the Labor Commissioner and Attorney General. It would further amend Labor Law section 218 to make violations of the new article subject to the department’s existing civil penalty and compliance-order framework, thereby expanding state labor enforcement authority over algorithmic hiring and workplace decision systems.
The bill’s tone and structure indicate strong support for worker privacy, transparency, and anti-discrimination protections in the use of workplace technology. The available record contains no committee transcript or vote history, so there is no direct evidence of opposition or amendment debate in the provided materials. Based on the text alone, the measure appears designed to address concerns about bias, hidden data use, and overreliance on automated systems, suggesting a generally reform-oriented and protective legislative posture.
The likely areas of contention are the bill’s regulatory reach and compliance burden. Employers and vendors may object to mandatory annual audits, public registry submissions, multilingual notice requirements, data-retention obligations, and restrictions on using automated outputs without meaningful human oversight. Supporters are likely to emphasize that these requirements are necessary to prevent discrimination, protect privacy, and ensure workers and applicants can understand and challenge automated decisions. The bill also raises potential disputes over what counts as an automated employment decision tool, what constitutes a disparate impact, and how much employee data may be collected or shared during assessments.