Connecticut Employers Need to Prepare for New Workplace AI Law: 4 Items For Your To-Do List

Connecticut Employers Need to Prepare for New Workplace AI Law: 4 Items For Your To-Do List

Connecticut just enacted one of the most comprehensive artificial intelligence laws in the country, and employers need to start preparing now. The Connecticut General Assembly recently passed Senate Bill 5, now formally known as the Connecticut Artificial Intelligence Responsibility and Transparency Act, and Governor Ned Lamont has confirmed he will sign it. While the law covers a broad range of AI-related topics, several provisions cut directly to the heart of how employers hire, manage, and make decisions about their workforce. Some key obligations will kick in on October 1, while additional measures take effect in October 2027. Here’s what you need to know and four important items you should add to your to-do list.

The Big Picture: A Law Built Around Employment

Connecticut’s SB 5 is not a broad AI governance statute but rather a set of separate AI-related provisions linked together in a single 67-page law. For employers, the most significant portions regulate what the law calls “automated employment-related decision technology,” a term defined broadly enough to capture much of what HR professionals use today.

This definition could apply to many commonly used AI tools, including third-party hiring platforms, resume screening software, assessment tools, scheduling algorithms, and performance analytics systems. Routine software like word processors, spreadsheets, and email are excluded.

The law uses staggered effective dates, so compliance obligations will roll in over the next two years.

• The first key date – October 1, 2026 – triggers the anti-discrimination amendments, the developer-deployer framework, and a new WARN Act disclosure requirement.
• The second major date – October 1, 2027 – is when the interactive disclosure and pre-decision notice obligations kick in for automated employment decision tools.

Key Aspects Of The Law

When the new law takes effect, there are four things you must do:

Disclose when AI is involved in decisions.

Beginning October 1, 2027, if you deploy an automated employment decision tool that is intended to interact with applicants or employees, you must inform those individuals in plain language that they are interacting with the technology (unless a reasonable person would find that fact obvious).

The law’s notice requirements go further when AI plays a “meaningful” role in an employment outcome. When the tool generates output that serves as a substantial factor in a decision, you must provide a written pre-decision notice identifying that an automated tool is being used, the purpose of the tool and the type of decision involved, the trade name of the technology, the categories and sources of personal data analyzed, how that data is assessed, as well as your contact information.

Divide responsibilities with your vendors.

The law draws a line between “developers” (the companies that build these tools) and
“deployers,” which in most cases means you as the employer.

• Developers must provide deployers with sufficient information to enable compliance with the law’s requirements, where the technology is marketed or intended to materially influence employment decisions.
• Developers and deployers may also contract for the developer to assume the deployer’s notice obligations, but those allocations must be explicit.

Protect trade secrets, but be transparent about doing so.

The law includes a trade-secret safe harbor: if information is withheld on that basis, you must provide a notice identifying what is being withheld and the legal basis for withholding it.

Disclose AI-related layoffs in WARN Act notices.

Effective this October 1, any employer filing WARN Act notices must disclose to the Connecticut Department of Labor whether layoffs are related to the employer’s use of AI or another technological change. This makes Connecticut one of the first states in the country to require employers to identify any connection between workforce reductions and AI adoption.  

Using AI Is Not a Defense to Discrimination

Perhaps the most significant change for HR and employment counsel is a direct amendment to Connecticut’s anti-discrimination framework. The law amends the Connecticut Fair Employment Practices Act to state explicitly that the use of an automated employment-related decision technology is not a defense to a complaint alleging a discriminatory employment practice. In other words, you cannot escape liability for discriminatory outcomes simply by pointing to the algorithm.

That said, employers who take proactive steps will be better positioned if a claim arises. The Connecticut Commission on Human Rights and Opportunities or a court may consider evidence of anti-bias testing or similar proactive efforts to avoid discrimination. That includes the quality, efficacy, recency, scope, results, and the employer’s response to those testing results. This is not a safe harbor, but it is a meaningful incentive to invest in rigorous, documented bias testing before deploying these tools.

Additional Non-Workplace Provisions

While the employment provisions will be on the top of most employers’ minds, SB 5 covers considerably more ground.

• The law imposes transparency obligations on developers of AI systems capable of generating synthetic digital content (including AI-generated audio, images, text, and video) requiring that such content be marked and detectable as such by October 1, 2027.
• It also establishes what may be the most restrictive AI companion chatbot regulations in the country, including an extensive set of child-specific prohibitions.
• A separate provision creates whistleblower protections for employees of frontier AI model developers who disclose potential catastrophic risks.
• And the law establishes a state-managed regulatory sandbox allowing companies to test new AI technologies and products, a provision that drew support from the governor’s office.

What Employers Should Do Now

With the first compliance deadline less than six months away, the time to act is now. Here are four priority steps:

1. Inventory your tools. Map every automated system that interacts with your employees, from sourcing and screening to performance management and discipline. Evaluate whether any system’s output constitutes a “substantial factor” in a covered employment decision.

2. Review your vendor contracts. Identify which vendors are providing automated decision tools and begin negotiating responsibility allocations for notice and disclosure obligations. Confirm that developers can provide the information you’ll need to satisfy your deployer duties.

3. Build your bias testing program. Work with your FP attorney to obtain bias-testing results, methodologies, and update schedules. Document corrective actions taken in response to testing findings. While this doesn’t create a legal safe harbor, it may matter significantly in any enforcement action or discrimination claim.

4. Update your WARN Act protocols. Evaluate whether AI adoption or other technological changes contributed to planned workforce reductions, and document your reasoning for any characterization you make in a WARN notice.

Conclusion

We will continue to monitor the situation and provide updates as they unfold. Make sure you are subscribed to Fisher Phillips’ Insight System to receive the most up-to-date information directly to your inbox. For more information, contact your Fisher Phillips attorney, the authors of this Insight, or any attorney in our AI, Data, and Analytics Practice Group.