A bill headed to the Connecticut Senate floor would reshape how employers use artificial intelligence (AI) in the workplace — adding new regulatory requirements, compliance costs, and, in the public sector, a larger role for unions in decisions about technology.
Senate Bill 435 establishes a broad framework governing the use of AI in hiring, scheduling, performance evaluations, and workplace monitoring. The bill is now on the Senate calendar for potential action.
Under the proposal, employers would be required to notify employees when AI systems are used, conduct annual bias audits, provide explanations for certain employment decisions, and offer workers the ability to seek human review of those decisions. It also creates a private right of action, allowing employees and their unions to bring claims in Superior Court.
The bill’s scope is expansive. It applies to any employer “doing business in the state,” meaning both public and private employers would face these requirements.
AI and Collective Bargaining
Beyond transparency and oversight, the bill introduces a structural change: it would bring the use of AI directly into the collective bargaining process for public-sector employers.
Under SB 435, decisions involving AI systems, including those used for hiring, scheduling, or monitoring, would become mandatory subjects of bargaining. State agencies and municipal governments would be required to negotiate with unions before adopting or modifying these technologies.
This represents a shift in how workplace technology decisions are made, placing them alongside traditional bargaining topics such as wages, hours, and working conditions.
Union testimony in support of the bill reflects this focus.
AFSCME Council 4 Deputy Director Zak Leavy praised provisions that “ensure that AI cannot be used to reduce wages, fringe benefits, or nonovertime hours, and cannot be used to assume the duties and functions of bargaining unit members.”
Similarly, Connecticut AFL-CIO President Ed Hawthorne highlighted that the bill “makes the use of AI technology a mandatory subject of bargaining for municipal and state employees” and “prohibits employers from using technology to erode wages, reduce hours, or assume bargaining unit work.”
Other union testimony emphasized the same goal from a different angle. CSEA SEIU Local 2001 President Travis Woodward argued that “workers must have a voice in how these tools are implemented, including through collective bargaining, to ensure technology supports rather than replaces them.”
Taken together, the testimony makes the priority clear: limiting AI isn’t just about oversight — it’s about protecting union jobs from being replaced by technology, jobs that define the size and strength of the unions themselves.
Costs and Implementation Challenges
The bill also carries a direct cost for taxpayers.
According to the official fiscal note, implementing SB 435 would require additional resources across multiple agencies. Estimated costs include approximately $422,000 annually for the Department of Administrative Services and more than $300,000 per year for the Department of Labor, along with associated fringe benefit expenses.
These costs are tied to requirements for oversight, including approving independent auditors and monitoring compliance. Because the bill applies to municipalities as well, local governments could face additionaladministrative and legal expenses.
Concerns from State Agencies
State agencies responsible for implementing the bill have raised operational concerns.
The Department of Administrative Services (DAS) noted that Section 14 would restrict agencies’ ability to use modern technology, noting that “AI components now appear in most commercial software and hardware,” meaning the bill could “significantly limit the technology the state can purchase.”
The agency cautioned the impact could extend to systems already in use, writing that the bill could “force the state to stop using a product that has already been purchased… until approved by the legislature.”
DAS also warned the proposal may “limit our ability to modernize systems, increase costs, and prevent the state from keeping pace with technology developments.”
The Office of Policy and Management (OPM) raised similar concerns, noting that existing collective bargaining agreements already provide “strong, enforceable protections” for employees and warning that new statutory requirements could add complexity without improving outcomes.
Labor Commissioner Danté Bartolomeo also noted that her agency currently lacks in-house expertise to carry out the bill’s requirements, stating the department “does not directly employ any AI subject matter experts” and would face “a significant fiscal impact” to comply.
Business and Industry Response
Business groups across multiple sectors have raised concerns about the bill’s scope and potential impact.
The Connecticut Business and Industry Association (CBIA) warned the proposal “creates a complex and burdensome regulatory framework that would make compliance extraordinarily difficult for employers, discourage innovation, and expose businesses to significant litigation risk.”
Local chambers echoed those concerns. The Connecticut River Valley Chamber of Commerce called the bill “unworkable and costly,” warning it would impose “significant compliance burdens” while discouraging investment and innovation.
Healthcare providers raised additional red flags. The Connecticut Hospital Association said the proposal would be “operationally unworkable” and could disrupt workforce management at a time when hospitals are already facing staffing shortages.
The Lumber Dealers Association of Connecticut warned the definition of automated decision systems is so broad it could apply to “many routine business technologies used by employers every day,” exposing companies to costly audits, compliance requirements, and litigation risk.
Across these responses, a consistent theme emerges: applying extensive regulatory requirements to widely used technologies could affect how employers hire, manage, and schedule employees, potentially increasing costs and slowing adoption.
A Broader Policy Question
SB 435 reflects a broader policy challenge: how to regulate emerging technologies in a way that protects workers while allowing organizations to adapt and innovate.
Artificial intelligence is becoming increasingly integrated into everyday business operations, often embedded in widely used software systems. Regulating its use raises complex questions about oversight, accountability, and flexibility.
Private employers would face new compliance requirements, audits, and the risk of lawsuits. Public agencies would face those same burdens, with an added layer: they would have to negotiate with unions before adopting or changing workplace technology.
Even the agencies expected to carry out the law are warning it could slow modernization, increase costs, and make it harder to use tools already built into everyday software.
The approach isn’t entirely new. Earlier this session, lawmakers briefly inserted union-related provisions into an unrelated online safety bill (SB 5) before stripping the language out. While that proposal ultimately changed, it reflects a broader pattern of expanding labor priorities into legislation far beyond traditional workplace policy.
As lawmakers consider the proposal, the central issue is not whether AI should be governed, but how to strike a balance between oversight and practicality in a rapidly evolving technological landscape.
