Connecticut's AI Law Is Signed: What Professional Services Firms Need to Do

Governor Ned Lamont signed Connecticut's AI Responsibility and Transparency Act (SB 5) on May 11, 2026, making Connecticut one of the most consequential states in the country for professional services firms navigating AI compliance. If your firm uses AI to make or inform employment decisions — or if you operate client-facing chatbots — you have until October 1, 2026 to get your house in order.

This is not a law aimed at tech companies. It applies directly to law firms, accounting firms, staffing agencies, and consulting practices that use AI for everyday operational decisions about their own employees and candidates.

Here is what the law requires, who it covers, and the three steps your firm should take this month.

---

What Connecticut SB 5 Does

The Connecticut AI Responsibility and Transparency Act is a 71-page omnibus law that covers four distinct areas. Not all of them apply to most professional services firms — but the employment provisions apply broadly.

Employment AI notification. Any employer that uses AI as a "substantial factor" in hiring, promotion, scheduling, performance evaluation, discipline, or discharge of Connecticut employees or applicants must:

• Notify affected individuals that AI was used in the decision
• Disclose what categories of data the AI processed and for what purpose
• Offer a meaningful opportunity for human review when the AI-influenced decision resulted in an adverse outcome

This is the provision most professional services firms need to act on first. The effective date is October 1, 2026.

Chatbot disclosure. Operators of AI-powered chat interfaces must disclose that users are interacting with AI, not a human. This applies at the point of interaction. If your firm's website or client portal has an AI chat assistant, intake bot, or automated scheduling tool, Connecticut users must be notified. The disclosure requirement is simple — a sentence at the start of the interaction — but it must be present.

Frontier AI model requirements. Developers of large-scale AI models (the kind that run inside ChatGPT or Claude) have additional obligations around documentation, risk assessment, and safety protocols. This tier does not apply to professional services firms using AI tools — it targets the companies that build the models themselves.

Synthetic content provenance. Large generative AI providers — those with more than one million monthly users — must embed machine-readable provenance data into AI-generated audio, images, and video. Again, this targets platforms, not the firms using them.

For the typical professional services firm, two provisions are operative: employment AI notification (October 1, 2026) and chatbot disclosure (staggered effective dates — confirm your specific category in the signed bill text).

---

Who Is Covered

The law applies to employers with Connecticut employees or applicants, chatbot operators with Connecticut users, and developers of frontier AI models. For professional services firms, the trigger is almost always the employment AI provision.

You are covered if:

• Your firm uses any software that incorporates AI to screen job applications, rank candidates, score assessments, evaluate performance, or assist in scheduling or workforce decisions
• Connecticut employees or job applicants are affected by those decisions
• The AI plays a "substantial factor" role — meaning it meaningfully influences the outcome, not just generates background data a human then ignores

You are likely not covered (for employment AI) if:

• Your firm is below the threshold for employer-size thresholds in the signed bill (confirm in final bill text — thresholds may apply)
• AI is used only to draft job postings or generate initial candidate pools, with no role in screening or ranking
• All AI-assisted review is incidental and a human makes the decision independently

The hard case for staffing firms: If AI screening tools are part of your core delivery — you use an ATS with AI ranking, a matching algorithm, or automated screening assessments to narrow candidate pools for client placements — you are almost certainly a covered deployer. The tool's AI may be presented as a "feature," not a decision-maker, but the legal standard is whether it is a substantial factor in outcomes. For staffing agencies, it usually is.

---

Key Dates You Need to Know

Obligation	Effective Date
AI employment decision notification	October 1, 2026
Human review option for adverse outcomes	October 1, 2026
Broader deployer obligations	October 1, 2027
Law signed by Governor Lamont	May 11, 2026

The gap between signing and the October 1, 2026 effective date is approximately five months. That is enough time to audit your tools and update your candidate/employee notification process — if you start in the next 30 days. It is not enough time to start in September.

---

What Law Firms Need to Do

Connecticut SB 5 does not regulate legal work product. AI used to draft motions, review discovery, analyze case law, or produce client deliverables is not within the scope of this law. The law's employment provisions apply to the firm as an employer — not to AI used in client engagements.

Where law firms are exposed:

• Lateral hiring: If your firm uses any AI-assisted screening for attorney or staff positions — applicant tracking systems with AI ranking, resume screening tools, skills assessment platforms — those uses are covered when Connecticut applicants or employees are involved
• Performance review cycles: AI tools that generate performance summaries, score attorneys on productivity metrics, or flag billing anomalies as performance factors are potentially covered for Connecticut staff
• Scheduling and staffing tools: AI-powered workforce scheduling tools used for paralegals, associates, or staff in Connecticut offices

What to do by October 1, 2026:

1. Ask your HR software vendors whether their tools use AI in employment decisions. Most will know. Get a written answer.
2. Update offer letters and rejection communications to include a short disclosure when AI was used: which tool, what role it played, what data it processed.
3. Document a process for human review when an applicant or employee requests reconsideration of an AI-influenced adverse outcome.

ABA Formal Opinion 512, which governs ethical AI use in client matters, is separate and unaffected by SB 5. That obligation runs in parallel.

---

What Accounting and Consulting Firms Need to Do

Accounting firms and consulting practices face the same exposure as law firms on the employment side: if you use AI in hiring or staff management in Connecticut, you are covered. The client-facing AI obligations — chatbot disclosure, frontier model requirements — are unlikely to apply to most firms in these sectors.

Accounting firms:

• Audit your hiring workflow. Do you use applicant tracking software with AI ranking? Karbon, Canopy, or similar practice management platforms may have AI features that touch staffing decisions — check the settings and vendor documentation.
• Update candidate notifications for open positions. A short paragraph in rejection emails noting AI-assisted screening suffices as initial disclosure. Full specificity (what tool, what data, what purpose) is required when the decision is adverse and the applicant asks.
• Advisory opportunity: Connecticut clients in HR-heavy industries — healthcare, financial services, staffing themselves — will have the same question about their own compliance. The firm that walks in with a clear answer has a consulting opportunity.

Consulting firms:

• Internal hiring and performance tools are the primary exposure. Project staffing algorithms, skills-matching tools for internal resource allocation, and AI-assisted performance feedback platforms all warrant a review.
• If you advise HR or workforce management clients, Connecticut SB 5 is a billable conversation starting today. Your clients face the same obligations you do — and they will need help.

---

What Staffing and Recruiting Agencies Need to Do

Staffing agencies are the most directly affected professional services category under Connecticut SB 5. If your core service involves placing candidates in roles for Connecticut clients, AI-assisted screening is almost certainly a "substantial factor" in those placements. The law reaches you.

Immediate priorities:

1. Audit every AI touchpoint in your candidate workflow. Resume screening, job matching scores, video interview analysis, skills assessments, background check flag scoring — each is a potential covered AI use. List every vendor. Ask each one in writing: does your product use AI to score, rank, filter, or flag candidates? Save the responses.

2. Update candidate communication templates. At rejection or adverse outcome, your outreach must include:

• A statement that AI was used in the screening or placement decision
• The categories of data the AI processed (resume content, assessment scores, skills match algorithms)
• The purpose of the AI's role (shortlisting, ranking, matching)
• How the candidate can request human review

One paragraph in an email template covers most of this. Draft it now.

3. Build a human review process. When a candidate requests reconsideration of an AI-influenced decision, you need a documented process for who handles it, what they review, and how the outcome is communicated. "We'll escalate to a recruiter" is a starting point — document it.

For more detail on the staffing-specific provisions, see Connecticut SB 5 AI Hiring Notification Law: What Staffing Firms Must Do and Connecticut SB 5 AI Employment Law: Staffing Firm Compliance Guide.

---

The Three Things to Do This Month

The compliance clock for Connecticut SB 5 starts at signing. October 1, 2026 is closer than it looks. Here is where to spend your next 30 days:

1. Audit your AI tools for employment decision involvement.

Pull up every software subscription your firm has. For each one, ask: does this tool use AI, and does that AI influence any decision about an employee or applicant? The answer for many practice management, ATS, and HR platforms is yes. Write it down. That list is your compliance inventory — and the first thing an enforcement inquiry will ask for.

2. Check vendor agreements for disclosure language.

Most enterprise AI vendors in the HR space have begun including language about their AI use in their data processing agreements. Review your vendor contracts for sections about automated processing, AI decision-making, or algorithmic tools. If those sections are missing, send a written inquiry to the vendor. You need written documentation of what role their AI plays in employment decisions before October 1.

3. Update offer and rejection communications to include AI notification if needed.

If any of your AI tools influence hiring decisions for Connecticut positions, update your candidate communication templates now. A three-sentence disclosure covers the legal requirement: (1) AI was used, (2) here is what it processed, (3) here is how to request human review. Your HR or operations team can draft this in an afternoon.

---

The firms that will struggle with this law are not the ones that use AI — it's the ones that don't know where their AI is. Connecticut SB 5 does not prohibit AI-assisted hiring. It requires transparency about it. If you cannot answer "which AI tools are involved in our hiring process and what data do they use," that is the gap to close before October 1, 2026.

---

Frequently Asked Questions

Does Connecticut SB 5 apply to professional services firms?

Yes, if your firm uses AI as a substantial factor in hiring, promotion, discipline, or discharge decisions affecting Connecticut employees or applicants. Law firms, accounting firms, staffing agencies, and consulting firms that use AI-assisted screening or scheduling tools for their own staff are covered. The law does not target legal work product or client-facing advisory work — it focuses on employment AI and chatbot disclosure.

What is the deadline to comply with Connecticut's AI law?

October 1, 2026 for AI employment decision notification requirements — firms using AI in hiring or employment decisions must notify affected applicants and employees by that date. Broader deployer obligations take effect October 1, 2027.

Does Connecticut SB 5 affect how law firms use AI for client work?

Primarily no. The law targets employment decisions, not legal work product. A law firm using AI to draft motions, review contracts, or conduct research is not covered by the employment provisions. However, if the same firm uses AI to screen attorney candidates, evaluate associate performance, or flag staffing patterns, those uses are covered. ABA Formal Opinion 512 governs ethical AI use in client work separately.

What must staffing agencies do under Connecticut's AI law?

Staffing and recruiting agencies are the most directly impacted professional services category. If you use AI as a substantial factor in candidate screening, ranking, or placement decisions for Connecticut clients or candidates, you must notify affected candidates that AI was used, offer a meaningful human review option, and maintain documentation of the AI tools and their role in each decision. October 1, 2026 is the compliance date.

How is Connecticut SB 5 different from Colorado's AI law?

Connecticut SB 5 focuses on employment AI and chatbot disclosure, with a narrower scope than Colorado's ADMT framework. Colorado's law (SB26-189) covers consequential decisions broadly — housing, financial services, healthcare, employment — and applies to firms using AI in any of those contexts. Connecticut's law primarily targets employment decisions and chatbot transparency. Both laws enforce through state attorneys general, not private plaintiffs, for most provisions.

---

The Crossing Report covers AI adoption for professional services firm owners every Monday. Subscribe free to get the weekly briefing — or go premium for implementation guides, tool comparisons, and state-by-state compliance updates.