BILL
SUBJECT
DATE
The Business Council of New York State opposes S.7263 (Gonzalez) / A.65450-A (Zaccaro), which would impose liability on proprietors of chatbots that provide substantive responses that, if given by a natural person, would constitute the unauthorized practice of medicine, law, or other licensed professions under New York law. While we support the underlying goal of preventing AI systems from fraudulently impersonating licensed professionals, the bill as drafted raised serious concerns that could produce unintended consequences well beyond that stated purpose.
Overboard Language ay Restrict Legitimate AI Use by Licensed Professionals
The bill's prohibition bars any chatbot response that "would constitute a crime" under New York's unauthorized practice of medicine, law, or engineering statutes. As written, this could prohibit licensed professionals themselves from using AI tools in their own practice. A physician using an AI assistant to draft clinical notes, or an attorney using AI to analyze case law, could trigger the same standard. This is contrary to the sponsors' publicly stated intent that the bill does not restrict licensed professionals from using AI tools in the course of their work. The bill should include explicit language exempting AI use by, or under the direct supervision of, duly licensed professionals.
Bill Text Is Broader Than the Sponsors' Stated Intent
The Senate sponsor has stated publicly that S.7263 targets chatbots that impersonate licensed professionals - not general-purpose AI tools that answer health or legal questions. The bill's text, however, uses a much broader standard: any response that "would constitute unauthorized practice." This sweeps in AI products that never claimed professional credentials, never held themselves out as a doctor or lawyer, and fully disclosed their AI status to users. The operative language should be rewritten to target impersonation specifically - meaning AI systems that affirmatively claim, imply, or represent that they are a licensed professional. This narrower standard matches the sponsors' stated intent and produced a more enforceable, targeted law.
The Disclaimer Rule Created Struct Liability for Honest, Disclosed AI Conduct
Section 2(b) provides that proprietor "may not waive or disclaim" liability merely by notifying users they are interacting with a non-human chatbot. The intention of stopping back actors from using a buried disclaimer as a legal shield is understandable. But as written, this provision means a chatbot that clearly discloses its AI status, explicitly states it is not a doctor or lawyer, and never claims professional credentials can still be held liable for providing health or legal information. That is not fraud or impersonation - it is strict liability for honest conduct. The fix is straightforward: operators who clearly disclose AI status and do not claim professional credentials should be protected, while those who use disclaimers to cover actual impersonation should not.
We respectfully request that the sponsors consider the following amendments:
- Add an explicit exemption for AI tools used by or under the direct supervision of licensed professionals.
- Revise the operative prohibition to target impersonation - defined as an AI system affirmatively representing or implying that it is a licensed professional - rather than the broader unauthorized practice standard.
- Revise §2(b) to distinguish between operators who genuinely disclose AI status and those who use disclaimers to cover actual impersonation. Operators who clearly tell users they are interacting with an AI, and never claim to be a licensed professional, should not face liability.
- Consider safe harbor language for AI products that identify themselves as AI, do not claim licensure, and direct users to consult a licensed professional for definitive advice.
The Business Council is committed to addressing AI impersonation and believes these amendments will accomplish the sponsors stated objective of preventing AI systems from fraudulently impersonating licensed professionals while preserving the benefits of responsible AI use in New York. However, as currently drafted, we must oppose S.7263 (Gonzalez) / A.65450-A (Zaccaro).