Do I Need Client Consent For Ai

ABA Formal Opinion 512 says clients must be informed about AI use "in appropriate contexts" -- and most contexts involving substantive legal work qualify. The opinion doesn't require written consent for every AI interaction, but it establishes that attorneys have a duty under Model Rule 1.4 (Communication) to keep clients reasonably informed about the means used to accomplish their legal objectives. When AI materially contributes to legal work product, the client needs to know.

The distinction between consent and notification matters. Consent implies the client can say no. Notification means you're informing them as a matter of professional duty. Most state bars that have addressed this land closer to notification -- you're telling the client you're using AI, not asking permission. But some situations, particularly those involving confidentiality risks, require affirmative consent before you proceed.

What ABA 512 Requires: The Communication Duty

ABA Formal Opinion 512 applies Model Rule 1.4 to AI use with two key requirements:

Inform clients about AI use in appropriate contexts. The opinion deliberately avoids a bright-line rule about when disclosure is required because the circumstances vary. But the guidance is clear: when AI tools play a meaningful role in the representation -- drafting briefs, analyzing case strategy, reviewing documents, conducting legal research -- the client should be informed.

Obtain informed consent when confidentiality is at risk. Model Rule 1.6 requires client consent before disclosing confidential information, and entering client data into AI tools constitutes disclosure to a third party. If you're using ChatGPT Free or Plus (where OpenAI may train on your inputs), you need informed consent under Rule 1.6(a) before entering any client information. Enterprise platforms with data protections reduce but don't eliminate this concern.

The practical reading of ABA 512 is this: notification is the floor for any substantive AI use, and affirmative consent is required whenever client confidential information enters an AI system.

State Bar Variations: Where the Rules Diverge

State bars are interpreting the communication duty differently, and the variations matter for multi-jurisdictional practice.

California: The State Bar's guidance leans toward informed consent rather than mere notification. California attorneys should obtain client agreement before using AI tools on their matters, particularly when confidential information is involved. The proposed Rule 8.3 would formalize disclosure obligations.

New York: NYC Bar guidance emphasizes transparency but stops short of requiring affirmative consent for all AI use. The focus is on ensuring clients understand how AI is being used and that the attorney remains responsible for all work product.

Florida: Bar Advisory Opinion 24-1 addressed AI in the billing context and reinforced that clients must understand what they're paying for. If AI reduces the time required for a task, the client needs to know that's why their bill looks different.

Texas: The state bar follows the ABA 512 framework closely. Notification is expected for substantive AI use; consent is required when confidential information is at risk.

Illinois: Has been among the more cautious jurisdictions, with bar guidance suggesting that AI use should be addressed in engagement letters and that clients should have the opportunity to object.

When Consent Is Required vs. When Notification Suffices

Affirmative consent required: - Entering client confidential information into any AI tool (Rule 1.6) - Using AI tools on platforms that may train on your inputs (Free/Plus tiers) - Using AI for tasks where the client has expressed preferences about technology use - Sharing client documents with AI platforms for analysis or review - Any situation where the engagement letter restricts technology use

Notification sufficient: - Using AI for general legal research that doesn't involve client-specific information - Using enterprise AI platforms with contractual data protections for client work - Drafting initial outlines or templates that will be substantially revised - Internal brainstorming and strategy development using AI - Grammar, formatting, and editing assistance

No notification typically needed: - Using AI for purely administrative tasks (scheduling, formatting) - Spell-checking and basic editing - General knowledge questions unrelated to the client's matter - Internal training and education on AI tools

Template Language for Client Communication

Here's language you can adapt for engagement letters and client notifications:

Engagement Letter Clause (Comprehensive): "In the course of our representation, we may use artificial intelligence tools to assist with legal research, document drafting, analysis, and other tasks. These tools include enterprise-grade platforms designed for legal work with appropriate data security protections. All AI-generated work product is reviewed, verified, and approved by a licensed attorney before use. Our use of AI tools does not change our professional obligations to you, including our duties of competence, confidentiality, and diligence. If you have questions about our AI use or wish to discuss limitations on AI tools in your matter, please contact us."

Matter-Specific Notification: "We want to inform you that we utilized [Tool Name], an AI-assisted legal research platform, in preparing [specific work product] for your matter. All output from the AI tool was independently verified by [attorney name] against primary legal sources. Our use of this tool helped us [benefit -- e.g., identify relevant case law more efficiently], and all work product reflects the professional judgment of our legal team."

Consent Request (for confidential data): "We would like to use [Tool Name] to [specific purpose] in your matter. This would involve entering [type of information] into the platform. [Tool Name] operates under [data protection terms -- e.g., SOC 2 compliance, no training on client data]. We believe this will [benefit]. We need your consent before proceeding. Please confirm your agreement or let us know if you have concerns."

Building AI Communication into Your Practice

The firms handling this best aren't treating AI disclosure as a one-off obligation -- they're building it into their standard client communication workflow.

Update your engagement letter template. Add a standard AI use clause to every new engagement letter. This establishes the baseline expectation and avoids the awkwardness of raising AI use mid-engagement.

Create a firm AI policy you can share. Clients -- especially sophisticated corporate clients -- want to know your firm has a thoughtful AI policy. A one-page summary of your AI use guidelines, verification protocols, and data protection practices is a competitive advantage.

Notify proactively, not reactively. Don't wait for the client to ask if you're using AI. Proactive notification builds trust and demonstrates professionalism. Reactive disclosure after a problem occurs looks like damage control.

Track state bar developments. The notification-vs-consent landscape is evolving quarterly. Assign someone at the firm to monitor state bar opinions and update your practices accordingly. What's sufficient today may be inadequate in six months.

The Bottom Line: ABA 512 requires client notification for substantive AI use and affirmative consent when confidential information enters AI systems -- and building this communication into your engagement letters now prevents compliance headaches later.