Connecticut’s AI Law Signals a New Phase of Employment AI Regulation
Connecticut has passed one of the broadest employment AI transparency laws in the country. The Connecticut Artificial Intelligence Responsibility and Transparency Act (Senate Bill 5) marks a major shift in how states regulate artificial intelligence in hiring and workplace decisions. This new employment AI regulation focuses on whether technology materially influences employment decisions and whether employers can explain and defend those systems under scrutiny.
For employers, the most impactful part of the law covers “automated employment-related decision technology.” This includes tools like resume screeners, candidate ranking systems, chatbot recruiting workflows, automated interview analysis, and productivity scoring tools. The law aims to ensure that when AI plays a significant role in hiring or promotion decisions, employers are transparent about it.
What Does “Automated Employment-Related Decision Technology” Mean?
Connecticut defines this term broadly. It covers any technology that processes personal data and uses computation to generate outputs—such as predictions, recommendations, rankings, or scores—that are a “substantial factor” in making or influencing employment decisions. A “substantial factor” means something that “meaningfully alters the outcome” of a decision.
This definition goes beyond fully automated systems. It also captures tools where human reviewers are still involved but the technology materially shapes the outcome. For example, if a hiring manager uses an AI tool to rank job applicants, and that ranking strongly influences who gets an interview, the tool is covered under the law.
Examples of Covered Technologies
- Resume screening and ranking software
- Chatbots that interact with applicants
- Automated video interview analysis
- Productivity or performance scoring systems
- Workflow tools that generate recommendations for human reviewers
Connecticut’s approach mirrors a growing trend among states. Colorado’s revised AI law focuses on whether automated systems materially influence consequential decisions. California’s Civil Rights Council regulations apply existing anti-discrimination principles to automated decision systems. Texas’s Responsible Artificial Intelligence Governance Act takes a lighter-touch approach. Together, these laws show that states are increasingly scrutinizing how technology shapes employment outcomes.
New Disclosure Obligations for Employers
Starting October 1, 2027, Connecticut employers must provide two types of disclosures when using covered AI technologies:
1. Real-Time Interaction Disclosure
Employers must tell applicants or employees when they are interacting with automated employment-related decision technology—unless it would be obvious to a reasonable person (like a chatbot clearly labeled as such).
2. Pre-Decision Written Notice
Before a covered decision is made, employers must provide written notice that includes:
- That the technology is being used
- The purpose of the system
- The nature of the employment decision
- The trade name of the technology
- The categories and sources of personal data analyzed
- How the data is assessed
- Contact information for the employer
These requirements may be challenging for many employers. Most organizations buy recruiting and hiring tools from third-party vendors. They may not have enough information to explain how the system evaluates data, what data sources it uses, or how much influence it has on the final decision. This creates pressure on both vendors and employers to coordinate better on governance, documentation, and disclosure practices.
Vendor and Employer Responsibilities Are Converging
One notable aspect of Connecticut’s law is how it treats both developers and deployers of AI systems. Developers must provide deployers (employers) with the information needed to comply with the law. However, the law also allows developers and deployers to contractually allocate compliance responsibilities between themselves.
This reflects a practical reality: employment AI governance is no longer just an HR or legal issue. It now touches procurement, vendor management, contracting, compliance, privacy, information security, and litigation risk management. As more states adopt similar laws, employers will likely demand more from vendors—including disclosure-ready documentation, bias testing information, data mapping details, and clear explanations of how systems work in hiring workflows.
Employer Accountability Without Mandatory Bias Audits
Unlike New York City’s Local Law 144, Connecticut’s law does not require annual bias audits. However, it still sends a strong signal that proactive governance matters. Connecticut amended its anti-discrimination laws to state that using automated employment-related decision technology “shall not be a defense” to discrimination claims.
The law also allows courts and agencies to consider evidence of anti-bias testing and other proactive efforts to avoid discrimination. This includes the quality, scope, recency, and efficacy of those efforts, the test results themselves, and the employer’s response to identified issues.
Connecticut’s approach emphasizes transparency, disclosures, and governance without prescriptive audit mandates. But it creates practical incentives for employers to evaluate and monitor the technologies influencing employment decisions. The law reinforces that employers remain responsible for employment decisions—even when technology helps shape them.
AI-Related Layoff Disclosures Expand the Law’s Reach
Connecticut’s law also reaches beyond hiring. Starting October 1, 2026, employers who submit WARN notices for covered layoffs or workforce reductions must disclose whether the layoffs are related to artificial intelligence or technological change.
This provision extends AI governance into workforce restructuring decisions. Regulators are interested not only in how employers use AI during recruiting and hiring, but also how technological change may shape layoff decisions. This may create documentation challenges for employers trying to determine when AI adoption materially contributed to staffing reductions.
Enforcement and Legal Exposure
The law relies primarily on the Connecticut Attorney General for enforcement through the state’s unfair trade practices framework. It does not create a private right of action—meaning individuals cannot sue directly under the statute. There is also a temporary cure period through the end of 2027 for the employment AI provisions.
That runway gives employers some flexibility during early implementation. However, organizations may still face litigation exposure through discrimination claims, regulatory investigations, and other employment-related legal theories. In many ways, the law reinforces existing principles rather than creating entirely new liability theories.
The broader legislation also includes provisions on generative AI transparency, AI companion systems, frontier model governance, and AI workforce development initiatives. This reflects how states increasingly view AI governance as a cross-functional policy issue—not just a narrow technology concern.
What Employers Should Do Now
Connecticut’s law will likely accelerate conversations many organizations are already having internally. Employers should consider:
- Inventorying all technologies used throughout hiring and employment workflows
- Identifying where automated systems materially influence decisions
- Reviewing vendor contracts and disclosure obligations
- Evaluating governance, testing, and documentation practices
Organizations may also need to rethink how they document human involvement in hiring workflows. As states continue moving toward “material influence” and “substantial factor” frameworks, the distinction between fully automated systems and human-assisted workflows may become less important than whether technology meaningfully shaped the outcome of a decision.
Connecticut’s law signals that employment AI regulation is entering a more operational phase. Employers need direct visibility into how technology influences employment decisions and how those systems operate in practice. The central question is increasingly: Does your organization understand how technology affects employment decisions, and can you explain and defend those systems under regulatory, litigation, or public scrutiny?
Comment