State rules on lawyers' use of AI

Your state's rules and ethics opinions control how you may use AI. ABA Formal Opinion 512 is the national baseline, but it interprets the Model Rules, which bind no one until a state adopts them. Most state guidance below is advisory ethics guidance; court-issued documents are noted as such.

This page covers the states whose documents we checked against an official source. Where a detail could not be confirmed, we say so or leave it out.

How do I find my state's rule on AI?

Start with your state bar's ethics opinions page and your state supreme court's notices to the bar, and search for "artificial intelligence". Then check the local rules and the assigned judge's standing orders for any court you file in (see Sanctions and court rules).

For a cross-state overview, Justia keeps a free AI and Attorney Ethics Rules: 50-State Survey. Trackers do not agree on how many states have issued guidance, so always confirm against your bar's own document.

California

  • Document: State Bar of California, Standing Committee on Professional Responsibility and Conduct (COPRAC), Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law
  • Date: 2026 version. It says: "This 2026 Practical Guidance replaces the 2023 version".
  • Official link: calbar.ca.gov

Key requirements:

  • Do not put confidential client information into a tool that may present material risks to confidentiality without the client's informed consent. Read the vendor's terms: "Reasonable efforts require more than reliance on generalized marketing assurances."
  • Review every output, including analysis and citations, before anything goes to a court.
  • Agentic AI systems may not make substantive legal decisions, give advice or file papers without meaningful lawyer supervision. The more autonomy a system has, the more oversight the lawyer owes.
  • Hourly bills must reflect time actually spent. General AI tool costs are overhead. Matter-specific AI costs can be billed only if the fee agreement says so, at actual cost, with no markup unless the client gives informed written consent.
  • A lawyer "must consider disclosure" to the client of intended AI use, and should check court rules that require disclosure to the tribunal.

"As a general matter, a lawyer must not input any confidential information of the client into a generative AI solution that may present material risks to confidentiality or security, absent informed client consent"

Florida

  • Document: The Florida Bar, Ethics Opinion 24-1
  • Date: 2024. No later opinion replacing it was found.
  • Official link: floridabar.org

Key requirements:

  • Research the tool's data retention and sharing policies. Informed client consent is recommended before confidential information goes to a third-party AI program.
  • Check AI research and drafts for accuracy and sufficiency.
  • Have firm policies, and treat AI output like work by a nonlawyer assistant.
  • Do not bill for time not worked or for learning the tool. Actual, reasonable AI costs may be charged only if the client is told in advance.
  • Client-facing chatbots must say clearly that they are AI, not a lawyer or staff member.

Read the full opinion at the link above for its exact wording.

New York

We found no statewide formal ethics opinion in New York, from the New York State Bar Association, as of November 2026. The New York City Bar has issued one.

  • Document: New York City Bar Association, Committee on Professional Ethics, Formal Opinion 2024-5, "Generative AI in the Practice of Law"
  • Date: August 7, 2024
  • Official link: nycbar.org

Key requirements:

  • Client consent is needed before using tools that share client information with third parties, and the consent must be informed.
  • Review AI output and the authorities it cites.
  • The duty to supervise lawyers and nonlawyers applies to AI use.
  • Routine embedded AI, as in Word, Westlaw, Lexis and search engines, does not need to be disclosed.

"Lawyers may wish to obtain advance client consent to use Generative AI that will involve sharing of client information, but, because such consent must be knowing, the client must understand the potential consequences"

New York courts may have their own rules on AI. Check the court system's rules before filing.

Texas

  • Document: Professional Ethics Committee for the State Bar of Texas, Opinion 705
  • Date: February 2025
  • Official link: legalethicstexas.com (the Texas Center for Legal Ethics publishes the Committee's opinions)

Key requirements:

  • Understand the technology reasonably well before using it.
  • If you are not reasonably sure the tool keeps information confidential, do not enter confidential information without consulting the client and getting consent.
  • You answer for what you file, however it was produced. Unchecked output risks breaching the rules on competence, meritorious claims and candor to the tribunal.
  • Do not bill hourly for time the AI saved. Reasonable direct AI costs may be passed on if the client agrees.
  • The Committee notes that the ABA and Florida suggest informed consent before using these tools, but it does not make consent mandatory in Texas.

"A lawyer may not, however, charge hourly fees for the time that was 'saved' by using the generative AI program."

New Jersey

  • Document: Supreme Court of New Jersey, Notice to the Bar, "Legal Practice: Preliminary Guidelines on the Use of Artificial Intelligence by New Jersey Lawyers" (court-issued)
  • Date: January 24, 2024. A follow-up Notice to the Bar, "Responsible Use of Artificial Intelligence (AI) and Related Technologies - Benefits of AI Policies" (March 30, 2026), encourages firms to adopt AI policies and attaches a template. It does not say it replaces the 2024 guidelines.
  • Official links: 2024 guidelines and 2026 notice

Key requirements:

  • Check and verify all AI-generated information.
  • Routine disclosure to clients is not required. Tell the client if the client asks, or if the client cannot make an informed decision without knowing.
  • Make sure an AI system is secure before entering any non-public client information.
  • Supervisory duties cover AI use by other lawyers and staff. The 2026 notice says firm policies are encouraged and are "not a 'safe harbor.'"

"Because AI can generate false information, a lawyer has an ethical duty to check and verify all information generated by AI to ensure that it is accurate."

Pennsylvania

  • Document: Pennsylvania Bar Association Committee on Legal Ethics and Professional Responsibility and Philadelphia Bar Association Professional Guidance Committee, Joint Formal Opinion 2024-200, "Ethical Issues Regarding the Use of Artificial Intelligence"
  • Date: 2024. The exact date is not shown on the copy we read.
  • Link: We could not find an official PBA or Philadelphia Bar copy. This link is a LawNext copy of the opinion, and the exact date is not shown on it. The Philadelphia Bar has published a summary.

Key requirements:

  • Check every citation and the material cited.
  • Tell clients about the benefits, risks and limits of AI, and seek their informed consent.
  • Do not put client information into AI that lacks adequate safeguards.
  • Make sure AI tools and assistants meet the lawyer's own obligations.
  • AI expenses must be reasonable and disclosed. Be open with clients and courts about AI use and its limits.

"Rule 1.4 requires the lawyer to inform the client of the benefits, risks, and limits of the use of generative AI."

Illinois

  • Document: Supreme Court of Illinois, Policy on Artificial Intelligence (court-issued)
  • Date: Announced December 18, 2024
  • Official link: illinoiscourts.gov announcement

Key requirements:

  • AI use by lawyers and litigants is authorized if it complies with legal and ethical standards.
  • Review AI-generated content thoroughly before filing it.
  • Lawyers answer for their final work product.

"Attorneys, judges, and self-represented litigants are accountable for their final work product. All users must thoroughly review AI-generated content before submitting it in any court proceeding"

Michigan

We found no formal ethics opinion in Michigan on lawyers' use of AI as of November 2026. Its numbered opinion on AI is addressed to judges.

  • Document: State Bar of Michigan, Ethics Opinion JI-155 (judicial ethics)
  • Date: October 27, 2023
  • Official link: michbar.org

Key requirements: Judicial officers must stay competent with advancing technology, including AI, and must not hand decisions over to AI.

"Judicial use of AI must distinguish between using an AI application to decide and using AI to inform a decision."

North Carolina

  • Document: North Carolina State Bar, 2024 Formal Ethics Opinion 1, "Use of Artificial Intelligence in a Law Practice"
  • Date: Adopted November 1, 2024
  • Official link: ncbar.gov

Key requirements:

  • The lawyer is personally responsible for reviewing and relying on AI work. Vendor tools fall under the duty to supervise nonlawyer assistance (Rule 5.3).
  • Client information may go into a third-party tool only if the lawyer is satisfied it is secure enough.
  • Routine use such as research needs no disclosure. Giving substantive tasks to an AI tool is like outsourcing, and needs the client's advance informed consent.
  • Bill only for time actually spent. Flat fees and documented AI expenses are allowed if they are not clearly excessive.

"Lawyer may not bill a client for three hours of work when only one hour of work was actually experienced."

Kentucky

  • Document: Kentucky Bar Association, Ethics Opinion KBA E-457, "The Ethical Use of Artificial Intelligence in the Practice of Law"
  • Date: March 15, 2024
  • Official link: kybar.org

Key requirements:

  • No duty to disclose routine AI research. Disclosure is needed if the work is outsourced, if the client is charged for AI, or if court rules require it.
  • When AI cuts the lawyer's time, the fee should come down accordingly.
  • Charge AI costs only with the client's advance agreement, confirmed in writing. Never bill clients for AI training or for keeping up with AI developments.
  • Get client consent where AI use would put confidential information at risk.
  • Firms need AI policies and training.

"No, there is no ethical duty to disclose the rote use of AI generated research for a client's matter unless the work is being outsourced to a third party; the client is being charged for the cost of AI; and/or the disclosure of AI generated research is required by Court Rules."

District of Columbia

  • Document: District of Columbia Bar, Legal Ethics Opinion 388, "Attorneys' Use of Generative Artificial Intelligence in Client Matters"
  • Date: April 2024
  • Official link: dcbar.org

Key requirements:

  • Understand the tool. Run a test and verify output, especially citations.
  • Find out whether the provider or later users could see client information.
  • Firms must take steps to ensure that AI use, including by vendors, complies with the rules.
  • Without a prior agreement, do not charge for the "perceived value" of AI work. Out-of-pocket AI costs may be passed through if the client agrees.

"Will information I provide to the GAI be visible to the GAI provider or other strangers to the attorney-client relationship?"

Washington

  • Document: Washington State Bar Association, Committee on Professional Ethics, Advisory Opinion 202505, "Artificial Intelligence-Enabled Tools in Law Practice"
  • Date: 2025
  • Official links: ao.wsba.org and wsba.org PDF

Key requirements:

  • Understand consumer and legal AI tools well enough to use them within the rules. The lawyer, not the tool, is responsible for the work.
  • Vet a tool before entering client information, using the vendor-evaluation factors from the WSBA's cloud-computing opinion (AO 2215).
  • Time spent using the tool, such as writing prompts, may be billed. Time the tool saved may not.
  • The opinion also covers supervision (RPC 5.1 and 5.3), communication (RPC 1.4) and candor (RPC 3.3).

"they may not charge for the 'time saved' under RPC 1.5(a)."

Georgia

  • Document: State Bar of Georgia, Generative AI Toolkit, prepared by its AI committees with the Law Practice Management Program. It is expressly guidance, not a formal advisory opinion. We found no Georgia formal advisory opinion on AI.
  • Date: Last updated July 9, 2026
  • Official links: toolkit PDF and AI and emerging tech page

Key requirements:

  • Review and vet all AI output, and supervise it as you would a person's work.
  • Do not misrepresent AI use to clients.
  • Get informed consent before adopting AI tools that affect billing, and keep fees reasonable. A written engagement agreement is recommended.
  • Confidentiality is covered in its section on Rule 1.6.

"Ultimately, AI is a tool. It cannot apply legal reasoning, understand justice or assume ethical responsibility. That remains the role of the lawyer."

What if my state is not listed?

Other states have issued guidance that we have not yet checked against official sources, so we do not describe it here. Use the Justia survey above to find it, then read your bar's own document.


Last reviewed November 9, 2026. This is general information, not legal advice.

About Faradex. Faradex is a private AI workspace for small and mid-sized firms that handle confidential client work. Each firm gets its own Workspace on dedicated infrastructure in the United States. Files and conversations are held in memory and are gone when the Workspace is reset. AI model: Currently Opus via AWS Bedrock, more information is available on our website. AWS does not store prompts or responses, and model providers have no access to them. Pricing: from $120 a month, no seat minimum; see faradex.ai/pricing.