On October 31, 2026, the House of Delegates of the New York State Bar Association (NYSBA) approved the report Legal Ethics Guidelines for Lawyers' Use of Artificial Intelligence. The report comes from NYSBA's Committee on Artificial Intelligence and Emerging Technologies (AIETC), chaired by Vivian D. Wesson. Our managing member, Luca CM Melchionna, chaired the Committee's Ethics Workstream, and Filippo Maccaferri was among the reviewers of the draft.
The full report is available on NYSBA's website.
Why New Guidance Now
NYSBA first addressed AI in April 2024, when the House of Delegates approved the report of its Task Force on Artificial Intelligence. That report laid the groundwork, and the new guidelines build on it.
A lot has changed in two years. Lawyers now use AI to draft, research, review documents and take notes on calls. Some tools can act on their own, sending messages or changing records with little human input. Courts across the country have sanctioned lawyers for filing briefs with made-up citations, and New York's appellate courts have now done the same. The updated guidelines respond to this new reality.
The Rules Have Not Changed
The main point of the report is simple. The New York Rules of Professional Conduct apply the same way whether or not a lawyer uses AI. The guidelines do not amend the Rules or create new disciplinary obligations. They explain how existing duties, such as competence, confidentiality and candor to the court, apply when AI is part of the work.
The guidelines also make clear that no lawyer is required to use AI. A lawyer who chooses to use it, though, stays responsible for the result.
What the Guidelines Cover
Verification comes first. A lawyer must check AI output before relying on it, and before filing anything in court must confirm that every cited case and quotation exists and says what the brief claims it says.
Other guidelines deal with everyday questions. Lawyers need to understand where client information goes once it is entered into an AI tool. They must tell clients about AI use in certain situations, for example when the client asks or when AI output is a significant basis for advice on a decision such as whether to settle. On billing, a lawyer may charge for time actually spent, including time spent checking AI work, but not for time spent learning how to use a tool, and the cost of the tool itself is generally overhead.
The report also addresses AI notetakers that record and summarize conversations, supervision of staff and vendors who use AI, and AI systems that can take action on their own. For those systems, a lawyer should approve consequential steps, such as filing a document or sending substantive information to a client or a court, before they happen.
What Comes Next
The Committee recommends reviewing the guidelines regularly, especially if the law or the technology changes significantly. AI tools will keep evolving, and the guidance is meant to evolve with them.