Last Updated on Feb 19, 2026 by Kurt Dunphy

State Bar Rules on AI Use: What Lawyers Need to Know About AI Compliance

A solo practitioner drafts a motion using AI but skips the final review. The brief contains fabricated legal citations, and they now face disciplinary proceedings and a formal bar complaint. This scenario is no longer hypothetical. It is a real risk attorneys face in a high-tech landscape.

State bar rules on AI use provide the necessary framework for ethical adoption of generative AI platforms. These rules cover attorney competence, client confidentiality, and professional responsibility. As legal technology capabilities evolve, bar associations are issuing ethics guidance to clarify how existing rules apply.

To maintain professional integrity, lawyers must comply with these rules and standards. This article explores state-specific guidance and core ethical duties. You will learn how to supervise AI tools effectively, validate AI-generated output, and protect work product through the appropriate due diligence procedures.

Key Takeaways

Human oversight of AI use is mandatory. Attorneys must verify all AI-generated output before it is included in court filings.
Confidentiality requires risk-aware selection of legal technology software that meets industry encryption standards and ensures data isolation. "Enterprise" or "Legal-Grade" AI is often required to ensure a private instance or isolated data architecture.
State bar rules on AI use may vary by jurisdiction-specific standards, but the ABA Model Rules of Professional Conduct set a universal baseline.

Comparing State Bar Rules on AI Use Across Jurisdictions

Not all states have formal state bar rules regarding AI use. However, many ethics committees have issued opinions or state bar AI guidance. California, Florida, and New York currently lead the way with detailed ethical guidance on the safe adoption of AI in legal practice.

As of early 2026, several states have moved beyond general statements to provide specific frameworks for attorneys:

California: The Bar associations published a "Practical Guide" emphasizing that attorney competence requires an understanding of large language models before use, including the risks of "hallucinations" and data privacy.
Florida: Opinion 24-1 mandates that attorneys disclose AI use when it impacts client billing or costs.
Texas: Opinion 705 (February 2025) clarifies that legal practice involving AI requires human oversight of AI-generated work to prevent the submission of fabricated case citations.
New York: Formal Opinion 2025-6 focuses on the professional responsibility of using AI to record and transcribe client meetings, emphasizing confidentiality and consent.
North Carolina & Oregon: Both states issued ethics guidance (including North Carolina's 2024 Formal Ethics Opinion 1) that requires lawyers to maintain competence in legal technology and vet legal technology vendors thoroughly.
Pennsylvania: Joint Formal Opinion 2024-200 warns that while AI can generate work, it does not replace the professional duty to verify all case law references.

Other states offer informal guidance through bar publications and FAQs. While these serve as a risk-aware roadmap rather than a formal disciplinary code, they are often cited by courts to define the "standard of care" in malpractice cases.

In states with no specific guidance, attorneys must still comply with professional conduct standards. The ABA Formal Opinion 512 (released in 2024) serves as the primary "north star" in these jurisdictions, providing a verified framework for competence, confidentiality, and communication. Guidance is evolving quickly as bar regulators and judges respond to technology shifts.

What Are Lawyers' Core Ethical Obligations When Using AI?

New legal technologies do not change a lawyer's professional duties. Applicable duties include:

Duty of Competence (Rule 1.1)

Lawyers must maintain competence in legal technology under Comment 8 of Rule 1.1. They must understand the benefits and risks of hallucination with large language models. This means you must evaluate a tool's training data and validate its output. You cannot simply delegate tasks without knowing how machine learning works.

Confidentiality and Data Security (Rule 1.6)

Rule 1.6 protects client confidentiality through data security measures. Lawyers must protect client information and ensure that AI platforms do not use their data for training. It is important to verify that legal technology vendors have secure data protection frameworks, evaluate privacy policies, and SOC 2 certification before use.

Supervision of Non-Lawyer Assistance (Rule 5.3)

Rule 5.3 clarifies supervision requirements for non-lawyer assistance. Attorneys must supervise every draft and generate final versions through human review. You remain accountable for the work. If the AI makes a mistake, the disciplinary authorities will look at the lawyer. Professional responsibility cannot be delegated to AI.

Candor to the Court (Rule 3.3)

Rule 3.3 requires verification of all legal citations before filing documents with the court. Lawyers must verify every case-law reference and factual claim, as judges have imposed sanctions in cases such as Mata v. Avianca for AI-generated output containing fabricated cases. Some courts now require you to disclose if AI drafted the filing. Courts are now treating the failure to check AI-generated citations as a lack of candor toward the tribunal.

A significant development in late 2025/early 2026 is the Duty to Disclose to the Client. While Florida led with billing, other states are now debating whether lawyers must tell a client any time they use AI to draft a substantive motion, similar to how they might disclose the use of an outside contract lawyer.

What Tasks Can Lawyers Delegate to AI Under Current Rules?

Attorneys can delegate routine tasks to AI, provided they retain human oversight.

Generally Permitted AI Uses (With Supervision):

Legal research and case law analysis with citation verification.
Contract review and identifying potential risks.
Drafting an initial legal document to eliminate "blank page" syndrome.
Due diligence procedures and sorting large volumes of clients' data.

High-Risk AI Uses Requiring Extra Caution:

Generate court filings without a documented review process.
Providing legal advice directly via chatbots without attorney intervention.
Making algorithmic decision-making choices on litigation strategy.
Inputting confidential information into public AI models like the free version of ChatGPT.

How to Comply With State Bar Rules on AI Use

Use this framework to monitor your firm’s use of AI.

Before Adopting an AI Tool:

Evaluate the vendor’s data privacy policies and ensure they possess SOC 2 certification for secure data handling.
Confirm the tool does not use client data for machine learning or "self-learning" purposes. Law firms are expected to use Enterprise/Closed-Loop systems where data is contractually excluded from the model's training set.
Ensure the software aligns with your specific state bar rules.
Validate the tool on non-confidential files to understand its hallucination risks.

While Using AI in Your Practice:

Always verify every case-law reference and legal draft before it leaves your desk.
Maintain final decision-making authority over all court filings and documents.
Document AI involvement to comply with disclosure and quality control requirements.
Train staff on security requirements and monitor evolving state bar AI guidance.

Resources to Stay Informed on AI and Attorney Ethics

Building an attorney's competence is an ongoing responsibility. Use these resources to track evolving state bar rules on AI use.

Official Ethics Guidance:

ABA Formal Opinion 512: The definitive national baseline for generative AI ethics and Rule 1.1 compliance.
The Florida Bar AI Toolkit: A "one-stop" shop featuring disclosure templates and specific fee-sharing guidance.
NYSBA AI Task Force Report (2025): Provides a comprehensive, phased roadmap for secure AI adoption.

Education and Training:

"AI Ethics: Achieving Compliance" (PLI): A deep dive into hallucination risks, agentic AI, and vendor data policies.
"GenAI for the Legal Profession" (Berkeley Law Executive Education): A 2026 program focused on human-in-the-loop workflows.
"Ethics 2026: AI & the Duty to Supervise" (Barbri): Specialized training on Rule 5.3 regarding the supervision of AI "nonlawyers."

Conferences and Monitoring:

Legalweek (NYC) & ABA TECHSHOW (Chicago): The industry's premier annual events for hands-on legal technology software training.
Stanford HAI (Human-Centered AI): Follow their "Legal AI Index" for verified technical accuracy benchmarks, used by firms to satisfy due diligence requirements for measuring model reliability and hallucination rates.

How Spellbook Helps Lawyers Meet Bar Ethics Requirements

Spellbook is designed to help attorneys comply with professional conduct standards. Unlike public AI tools, Spellbook is SOC 2 certified and adheres to legal-grade encryption standards. Your confidential work always remains secure because Spellbook’s AI does not store or train on your data.

The platform mandates human oversight of AI-generated work to ensure attorneys maintain final decision-making authority. Key features include:

Review & Redlining: Scans contracts to identify and redline risky provisions and missing clauses.
Smart Clause Drafting: Suggests language from your firm's verified precedents.
Benchmarking: Validates document coverage against 2,300+ industry standards.
The "Associate" Agent: Automates multi-step workflows while keeping you in control.

Because Spellbook works inside Microsoft Word, where attorneys spend most of their time, it is exceptionally easy to use. Book a demo to explore how Spellbook helps you stay compliant while drafting 10x faster.

Frequently Asked Questions

Are Lawyers Required to Disclose AI Use to Clients?

Yes, in many circumstances. Rule 1.4 promotes transparency by requiring client disclosure of the "means" by which a client's objectives are accomplished. Disclosure is mandatory if the AI use involves sharing confidential data with a third-party vendor (Rule 1.6) or if it impacts the basis of your fees (Rule 1.5). Some state bar rules on AI use explicitly require it in engagement letters. Even if not mandated, disclosure is a professional best practice.

Can I Use ChatGPT for Legal Work Under Bar Rules?

You can, but it can be risky. The free/consumer version of ChatGPT is a public tool, and anything you type into it could be used in its training data. This creates a massive confidentiality problem. Most bar ethics rules for AI require much stricter data security than a public chatbot provides.

What Happens if I Submit an AI-Generated Brief with Fake Cases?

You could face sanctions or disciplinary proceedings. Rule 3.3 requires verification of all legal citations before filing. Judges have no tolerance for AI-generated output that lacks human verification of citations.

Do I Need Special Training to Use AI Tools Ethically?

Yes. Lawyers should maintain competence in legal technology. In states such as Florida and New York, bar associations now mandate CLE credits to ensure ethical technology adoption.

Which State Has the Strictest AI Rules for Lawyers?

No state is the "strictest". However, Florida and California have jurisdiction-specific ethics guidance. These states require human oversight of AI-generated work and clear disclosure rules. Always check your local bar associations.

Can AI Practice Law on Its Own Under Bar Rules?

No. This would be the unauthorized practice of law. AI is a tool that helps lawyers draft and review, but the lawyer must supervise the final work product.

‍