New York's AI Disclosure Bill Would Require Every Firm to Warn Clients That AI Outputs May Be Wrong

Most of the AI disclosure bills moving through state legislatures target chatbots and intake forms — the places where AI talks to your clients before you do. New York A 3411 targets something different: the work product itself.

The bill, passed by the New York legislature on March 9, 2026, and currently at Governor Hochul's desk, would require AI systems to disclose to users that generated outputs may be inaccurate. For professional services firms, the implication is direct: if you use AI to draft a contract, prepare a tax analysis, develop a consulting recommendation, or generate any client deliverable, the output needs to carry an accuracy warning.

This is not yet law. But it passed both chambers, it's sitting with the Governor, and the direction is clear. Firms with New York clients or New York operations that aren't already updating their engagement letters are about to be behind.

---

Summary

New York A 3411 passed the state legislature on March 9, 2026, and is at Governor Hochul's desk. The bill would require AI systems to warn users that generated outputs may be inaccurate — creating a statutory accuracy disclosure obligation for professional services firms delivering AI-assisted work product to New York clients. For law firms, this builds directly on ABA Opinion 512. For accounting and consulting firms, it creates a new layer of disclosure duty that most engagement letters don't yet address.

---

What the Bill Actually Says

NY A 3411 targets what lawyers and policymakers are calling "GenAI outputs" — any content generated by a generative AI system. The disclosure requirement is clear in intent: users must be informed that what they're looking at may not be accurate.

In a professional services context, that means:

• A law firm using AI to draft a contract or brief must inform the client that AI was involved and that the output may contain errors
• An accounting firm using AI to prepare financial analysis or tax research must disclose the same
• A consulting firm using AI to develop strategy recommendations or competitive analysis faces the same requirement
• A staffing firm using AI to generate job descriptions, candidate assessments, or client reports is covered

The bill's reach is not limited to New York firms. It covers AI-generated outputs delivered to New York clients — meaning a Philadelphia law firm sending AI-drafted documents to a Manhattan client has a New York compliance obligation.

---

The Two-Layer Problem for Law Firms

For lawyers, NY A 3411 doesn't arrive in a vacuum. ABA Opinion 512, issued in 2024, already established that lawyers have a professional duty to supervise AI use and disclose AI involvement to clients in appropriate contexts. Most state bars have adopted or are considering similar guidance.

NY A 3411 would add a second, statutory layer on top of that: not just a professional ethics obligation to disclose, but a state law obligation to warn about accuracy. The combination creates a clear requirement that engagement letters and client deliverables address both:

1. Disclosure of AI use (ABA Opinion 512 / professional conduct rules)
2. Accuracy warning for AI-generated content (NY A 3411, when signed)

Firms that updated their engagement letters for ABA Opinion 512 compliance may need a second pass to add the accuracy warning language NY A 3411 would require.

---

The Engagement Letter Gap

The practical problem is this: most engagement letters for professional services firms — law, accounting, consulting — were written before AI was a meaningful part of the workflow. They address retainer terms, scope, fees, and professional standards. They say nothing about how the work is prepared.

That gap is closing through professional conduct rules, state statutes, and client expectations simultaneously. Here's the pattern across three firm types:

Law firms: ABA Opinion 512 created a disclosure and supervision obligation. NY A 3411 would add an accuracy warning requirement. Bar discipline and malpractice exposure meet at the engagement letter.

Accounting firms: No AICPA equivalent of Opinion 512 yet, but state CPA societies are moving. NY A 3411 would create a statutory disclosure duty independent of professional ethics guidance — which means accounting firms can no longer wait for the AICPA to act.

Consulting, staffing, and marketing agencies: No professional conduct overlay, but client contract obligations and state AI statutes increasingly fill that space. NY A 3411 applies if you deliver AI-assisted work product to New York clients.

The engagement letter is the right place to address all of this — before a deliverable dispute, before a malpractice claim, before a regulatory inquiry.

---

Three Things to Do Before This Becomes Law

1. Update your engagement letter AI clause.

If you don't have one, add one now. The minimum viable clause should state: (a) that the firm uses AI tools in preparing some work product, (b) that AI-generated outputs are reviewed by a licensed professional before delivery, and (c) that the client may request clarification on how any specific deliverable was prepared. Add an accuracy acknowledgment: "AI-generated content may contain errors. All work product is reviewed by a licensed professional prior to delivery."

Have the language reviewed by qualified counsel for your jurisdiction. If you have New York clients, request review specifically for NY A 3411 alignment.

2. Identify which deliverables need a disclosure footer.

Not every email needs an AI accuracy warning. But any deliverable that relies substantially on AI-generated content — a contract draft, a tax analysis, a research memo, a candidate assessment — should carry one. A one-line footer is sufficient: "Portions of this document were prepared with AI assistance and reviewed by a licensed professional." Create a template for your most common deliverable types.

3. Document your AI review workflow.

The disclosure only holds up if there's actual human review behind it. For each AI-assisted workflow your firm uses, document who reviews AI output before it reaches a client, what the review covers, and how errors are caught. This documentation is your defense if a client challenges a deliverable under NY A 3411 or a professional conduct rule. It's also what state bar compliance auditors and E&O insurers are beginning to ask for.

---

The Pattern Forming Around You

NY A 3411 is the fourth major state AI bill to move in March 2026 alone:

• Oregon HB 4154 (chatbot disclosure, private right of action) — passed, now law
• New Hampshire SB 640 (AI prohibited from providing licensed professional services without meaningful human oversight) — passed committee
• Washington HB 1170 and HB 2225 (AI transparency and chatbot safety) — passed March 13
• New York A 3411 (GenAI output accuracy warnings) — passed legislature, at Governor's desk

The pattern is not accidental. State legislatures are moving faster than federal law on AI — and professional services firms are in the crosshairs of every one of these bills because clients are involved, professional duties apply, and the harm scenario is specific and demonstrable.

Waiting until a bill is signed to prepare is no longer a viable compliance strategy. The direction is clear. The engagement letter is where this lands. Update it now.

---

Related Reading

• ABA Opinion 512: The Three-Document AI Compliance Checklist for Law Firms
• Oregon Just Gave Your Clients the Right to Sue You Over Your AI Chatbot
• NH SB 640: AI Can't Provide Licensed Professional Services Without Meaningful Oversight
• The FTC Just Defined AI Deception — What Professional Services Firms Must Know
• AI Liability Is Now an E&O Insurance Question — 3 Things Your Carrier Is About to Ask