AI Ethics for Lawyers: Using AI Without Violating Rules 1.1, 1.6, and 5.3

The Three Rules That Govern AI in Your Practice

State bars have spent the last three years issuing AI ethics opinions. The thread running through almost all of them: your professional responsibility obligations don't change because AI is involved. They expand to include AI.

Three ABA Model Rules - adopted in some form by every U.S. jurisdiction - sit at the center of legal AI ethics:

• Rule 1.1: Competence
• Rule 1.6: Confidentiality
• Rule 5.3: Responsibilities Regarding Nonlawyer Assistance

This article explains how each applies to AI, what state bar opinions have said, and the specific practices that keep you safely on the right side of the line.

Rule 1.1: The Duty of Competence Now Includes AI

Comment 8 to Rule 1.1 has required attorneys to "keep abreast of changes in the law and its practice, including the benefits and risks associated with relevant technology" since 2012. In 2024, the ABA's Formal Opinion 512 made explicit what was already implicit: AI is now part of that technology competence requirement.

What this means practically:

• You don't have to use AI. You do have to understand it.
• If you do use AI, you must understand the model's strengths and limitations - including hallucination risks.
• "I didn't know AI hallucinates" is not a defense.

A baseline of competence for any AI-using attorney:

1. You understand that current AI models can invent plausible-looking case citations. (See AI hallucinations in legal work.)
2. You know the difference between consumer-tier and enterprise-tier AI tools.
3. You can articulate what the AI tool does with your input data.
4. You verify AI output before relying on it.

Rule 1.6: The Confidentiality Problem

Rule 1.6(a) prohibits revealing "information relating to the representation of a client" without informed consent. Rule 1.6(c) requires "reasonable efforts to prevent the inadvertent or unauthorized disclosure of, or unauthorized access to, information relating to the representation of a client."

How AI changes this:

When you paste client information into a consumer AI tool, you may be transmitting confidential information to a third-party vendor that uses it for training. State bars have repeatedly held that this can constitute disclosure under Rule 1.6.

The 2026 standard of care:

• Consumer tiers: Avoid for any privileged client information unless you have written informed consent and you've verified data handling.
• Enterprise tiers with BAAs: Generally acceptable. ChatGPT Team/Enterprise, Claude Team/Enterprise, Gemini Workspace all offer zero-data-retention configurations. For the comparison, see ChatGPT vs. Claude vs. Gemini for lawyers.
• Private deployments: Even safer. Services like Claude Connect host AI within infrastructure your firm controls.

Practical guardrails:

1. Sanitize inputs. Use placeholders ("Client A," "the Defendant") instead of real names when you can.
2. Use the right tier. Pay the $30/month for enterprise. It's nothing against billable rate.
3. Get informed consent. Update your engagement letters to disclose AI use.
4. Audit. Periodically review what's been entered into AI tools by your team.

Rule 5.3: Supervising AI Like a Nonlawyer Assistant

Rule 5.3 governs your responsibility for non-lawyer assistants - paralegals, secretaries, contractors. Multiple state bar opinions (including New York, Florida, California, and the ABA's 2024 opinion) have extended this principle to AI tools.

The framing: treat AI like an unsupervised paralegal. Your supervision obligation applies. The fact that the "paralegal" is software doesn't reduce your duty.

What supervised AI use looks like:

• Set clear guidelines for which tools attorneys can use for which tasks.
• Train attorneys on prompt engineering and verification workflows (see the lawyer's guide to prompt engineering).
• Build verification checkpoints into your workflow.
• Maintain audit trails of AI use, especially for filings and client communications.

If your firm has more than two attorneys, you should have a written AI use policy. Our AI training program builds firm-specific policies as part of every engagement.

The Disclosure Question

Several court rules and ethics opinions now address whether attorneys must disclose AI use:

• To courts: Some federal districts now require disclosure of generative AI use in filings.
• To clients: Most jurisdictions don't require routine disclosure, but several recent opinions strongly recommend disclosing AI use to clients, especially when it affects billing.
• To opposing counsel: Generally not required absent specific court rules.

Our recommendation: build AI disclosure into your engagement letters by default. It's a low-cost way to stay ahead of regulatory trends.

The Billing Question

Can you bill for AI time? It depends - and it's one of the most-asked questions we get. See our full analysis in should you bill AI time to clients. Short version: you generally cannot bill the same hour twice (once for AI and once for "review"), but value-based pricing and reasonable time entries for AI-assisted work that you supervise are typically fine.

The Verification Standard

Across all three rules, one principle runs through: the lawyer remains responsible.

• Competence: You're responsible for understanding the tool.
• Confidentiality: You're responsible for what enters it.
• Supervision: You're responsible for verifying what comes out.

This is actually liberating once you internalize it. You don't have to avoid AI to be ethical. You have to use it like a competent professional who supervises an assistant. That's a standard every lawyer already meets in other contexts.

A 10-Minute Ethics Checklist for Your Firm

Run this exercise this week:

1. List every AI tool currently used by anyone at the firm.
2. For each tool, identify the tier (consumer / business / enterprise).
3. For each tier, verify the data handling policy.
4. Document which tools are approved for which categories of work.
5. Update engagement letters to disclose AI use.
6. Build a one-page AI use policy and distribute it.

If this exercise feels overwhelming, that's a strong signal that your firm needs structured help. Our AI Operating System service builds these governance layers into firms in 60 to 90 days.

The Bigger Picture

The ethics rules are not designed to prevent you from using AI. They're designed to ensure that AI use doesn't compromise your obligations to clients, courts, and the profession. Used correctly, AI strengthens your ability to meet those obligations - by reducing errors, expanding your capacity, and lowering the cost of access to legal services.

Used incorrectly, it's a sanctions case waiting to happen. The difference is operational discipline.

If you want help building AI ethics infrastructure into your firm - written policies, training, audit workflows - see our AI training program or schedule a consult.