AI is now part of how legal work gets done at large firms. It summarizes depositions, reviews documents, drafts correspondence, supports legal research, and helps attorneys manage the workload that comes with a full docket. Most of this is good. And most clients, once they understand how AI is being used, agree.
But most clients don’t even realize it’s being used in the first place.
For managing partners, risk and compliance leaders, innovation officers, and practice group heads at large and enterprise firms, that gap is worth closing. We’re beyond questioning whether to disclose AI use. Now, the goal is to do it well, at scale, across hundreds of attorneys and thousands of matters, in a way that strengthens client relationships.
This guide covers what the ethics guidance says, when disclosure is required, what a strong disclosure looks like, and how large firms can build a repeatable approach that treats transparency as a professional standard.
Why AI Disclosure Matters for Large Law Firms
Client relationships at large firms are built on trust, and trust depends on clients knowing how their matters are being handled. More and more, AI is playing a meaningful role in that work: shaping analysis, informing decisions, or processing confidential information. And when that happens, clients have a legitimate interest in knowing.
And that interest is showing up in practice. Sophisticated corporate clients are now asking about AI governance in RFPs and outside counsel guidelines with increasing regularity. They want to know what tools are being used, how their data is handled, and whether human attorneys are reviewing AI-generated work. Firms that can answer those questions clearly, and can demonstrate a governance framework behind the answer, have an advantage in competitive pitches and client retention.
There's also a professional responsibility dimension that matters at the firm level. When disclosure practices are inconsistent — when one partner handles it differently than another, or when it happens at the matter level but not in engagement letters — the firm carries the risk. At large firms, where AI use spans many practice groups, dozens of offices, and thousands of active matters, ad hoc disclosure is an unnecessary exposure.
The good news? Firms that get ahead of this and make disclosure a standard part of how they engage clients tend to find it builds confidence, rather than eroding it.
What Ethics Guidance Says About Disclosing AI Use
The clearest professional responsibility guidance on AI disclosure comes from ABA Formal Opinion 512, issued July 29, 2024. It was the ABA's first major ethics opinion on the use of generative AI in legal practice, and it remains the central reference point for firms navigating disclosure obligations.
The opinion grounds AI disclosure obligations in several of the Model Rules:
Model Rules
| Model Rule | What It Requires | What It Means for AI Disclosure |
| Rule 1.1 — Competence | Lawyers must understand the tools they use well enough to use them effectively and responsibly | Attorneys using AI must understand how it works, where it can fail, and how to verify its output |
| Rule 1.4 — Communication | Lawyers must keep clients reasonably informed and explain matters enough for clients to make informed decisions | Disclosure may be required when AI use is material to how a matter is handled |
| Rule 1.6 — Confidentiality | Lawyers must take reasonable measures to prevent unauthorized disclosure of client information | Before using a GenAI tool, lawyers must assess whether it adequately protects client data, especially tools that retain or train on input |
| Rule 1.5 — Fees | Lawyers may only charge reasonable fees | How AI use is billed — charged as a fee, passed through as a cost, or absorbed as overhead — must be transparent |
| Rules 5.1 / 5.3 — Supervision | Supervising attorneys are responsible for ensuring associates and non-lawyers comply with the rules | Partners must ensure that AI use by their teams meets ethical standards firmwide |
One finding from Opinion 512 worth emphasizing for large firms: boilerplate consent language buried in an engagement letter is generally not considered sufficient for informed consent when it comes to AI use. Clients need to understand what they're consenting to, in language they can actually read and evaluate.
State bars and local courts are adding their own layer. California, Florida, New Jersey, New York (including the NYC Bar's Formal Opinion 2024-5), and Pennsylvania have all issued guidance or have active rulemaking underway. Some courts are requiring attorneys to certify their use of AI in filings.
Because the landscape is evolving rapidly and varies by jurisdiction, firms should confirm their specific obligations with their own ethics counsel. Be sure to build a disclosure framework flexible enough to adapt.
Boilerplate vs. Meaningful Disclosure
| Boilerplate (generally insufficient) | Meaningful disclosure |
| "The firm may use technology tools, including artificial intelligence, in the course of the representation." | "We use AI tools in [specific contexts: document review, deposition summarization, research support]. These tools are reviewed by a supervising attorney before any output is used in your matter. Your confidential information is not used to train third-party models." |
When Should a Law Firm Disclose AI Use to Clients?
There isn't a single trigger for disclosure — context matters. But there are situations where the obligation is clear, and large firms benefit from identifying those situations explicitly in their policies rather than leaving it to attorney judgment.
Decision Matrix
| Situation | Approach | Why |
| AI meaningfully informs a significant decision about the representation | Disclose | Client has a right to know how their matter is being handled |
| AI is used to process confidential client information in a self-learning or third-party tool | Get informed consent before use | Rule 1.6 requires reasonable measures to protect confidentiality |
| AI use affects how the client is billed | Disclose | Rule 1.5 requires billing transparency |
| Client asks, or outside counsel guidelines require disclosure | Disclose and respond specifically | Contractual and professional obligation |
| A court or local rule requires proactive disclosure | Disclose | Compliance obligation |
| AI tool is routine, embedded, and treated as overhead (e.g., spell check, citation formatting) | Generally not required | No material impact on representation or billing |
The last row is worth noting: not every use of AI requires disclosure. Embedded tools that function like infrastructure (and are billed like it) generally don't create a disclosure obligation.
The key question is whether AI use is material to the representation. Does it meaningfully affect how the work gets done, what the client pays, or what information is shared with external systems? If yes, disclose.
What to Include in an AI Disclosure
When disclosure is warranted, the content matters as much as the fact of disclosure. A strong AI disclosure covers six elements:
Anatomy of an AI Disclosure
- A plain-language explanation of how and where AI is used. Clients shouldn't need a technical background to understand it. "We use AI tools to summarize deposition transcripts and flag relevant passages for attorney review" is more useful than "We employ large language models in our legal operations."
- The categories of tools or tasks involved. Document review, drafting support, research, summarization, chronology building: clients benefit from knowing which parts of their matter AI touches.
- How client data is protected. Specifically, whether data is retained by third-party tools, whether it's used to train models, and what security and compliance standards apply. Filevine's security and compliance infrastructure is an example of what enterprise-grade data protection looks like in practice.
- The role of human oversight and attorney review. Clients want to know a licensed attorney is accountable for the work. Reassure them that AI is supporting the attorney's judgment, not replacing it.
- How AI is treated in billing. Is it charged as a separate fee, passed through as a cost, or absorbed as firm overhead? Clients increasingly ask this question directly, and firms should have a clear answer.
- The client's ability to ask questions or set preferences. Good disclosure is a conversation, not a one-way notice. Giving clients the ability to ask follow-up questions or express preferences about AI use, even if the answer is that certain tools are non-negotiable for efficiency, reflects good practice.
How Large Law Firms Can Disclose AI Use
There's no single channel that works for every situation. Large firms typically need a combination of approaches, applied appropriately depending on the context.
Disclosure Methods
| Method | What It Is | Best For | Watch-Outs |
| Engagement letter or addendum | Written disclosure at the start of the representation | Establishing baseline consent; addressing billing and data handling | Boilerplate isn't enough; must be specific and readable |
| Standalone AI use policy | A client-facing document that explains the firm's AI governance | Sophisticated clients; RFP responses; ongoing relationships | Needs to stay current as tools and practices evolve |
| Matter-intake or onboarding conversation | A discussion at the start of a matter | When AI use is significant to a specific case | Should be documented; not a substitute for written disclosure |
| RFP and pitch responses | Proactive disclosure in competitive processes | Corporate clients with outside counsel guidelines | An opportunity to differentiate; don't bury it in boilerplate |
| Matter-level discussion | Conversation when AI use is significant to a specific case | Complex matters where AI plays a substantive role | Supplement, don't replace, written disclosure |
For large firms, the engagement letter and standalone AI policy are the foundation. The other methods layer on top for situations that warrant more specific or more proactive disclosure.
Building a Firmwide AI Disclosure Framework
The disclosure challenge scales up with larger firms. It must be consistent across hundreds of attorneys, multiple offices, and thousands of active matters. That requires an entire framework, custom built for the practice.
A firmwide AI policy and governance structure
Before attorneys can disclose AI use consistently, the firm needs to define what AI use looks like. Find out which tools are approved, how they're categorized, what oversight is required, and how billing is handled.
Consistent disclosure starts with consistent policy.
Standardized templates that allow matter-level flexibility
Engagement letter language, AI use policy documents, and disclosure addenda should be standardized at the firm level. Individual attorneys shouldn't be writing their own from scratch.
At the same time, the templates need room for matter-specific details. What's used on a large commercial litigation matter differs from what's used on an employment matter.
Training attorneys and staff
Model Rules 5.1 and 5.3 put supervisory responsibility squarely on partners and senior attorneys. That responsibility includes ensuring that junior attorneys and staff who use AI understand the disclosure obligations and handle them consistently.
Disclosure training should be part of onboarding and updated as guidance evolves.
Documentation and recordkeeping
When a client gives consent, written or in documented conversation, that consent should be captured in the matter record. At scale, this requires systems that make recordkeeping straightforward.
Filevine's data and systems infrastructure and core workflow tools are built to support the kind of structured recordkeeping that makes this manageable firmwide.
A review cadence
The ethics landscape is constantly shifting. State bar guidance issued this year may be superseded or clarified next year. Firmwide AI disclosure policies should have a scheduled review — at least annually — to stay current.
Next steps
For firms looking to understand how AI capabilities fit into this framework, Filevine's four AI capabilities guide and Future of AI in Law white paper are useful starting points.
How to Address Common Client Concerns About AI Use
When clients ask about AI, they're usually asking one of four questions.
Firms that have clear, grounded answers build confidence. Firms that don't have answers, or whose answers are inconsistent across attorneys, leave room for doubt that erodes trust over time.
"Is my confidential information protected?"
Yes. The answer should be specific. Clients want to know whether their data stays within the firm's systems, whether any third-party tools retain input, and what security standards apply. The answer to this question is one of the strongest reasons to choose enterprise-grade, purpose-built legal platforms over general-purpose AI tools.
"Is a human still reviewing the work?"
Yes. Attorney oversight is an ethical obligation. Clients should understand that AI supports the attorney's judgment; it doesn't replace it. The specialized tools designed for legal practice are built with this oversight model in mind.
"Am I being billed for AI?"
This depends on how the firm has structured it, and the answer should be consistent with the engagement letter. Many firms treat routine AI use as overhead. When AI accelerates work significantly — and the client benefits from that efficiency — some firms pass through a cost or adjust their approach to billing accordingly. The key is that the client knows which approach applies before the bill arrives.
"How do you keep the work accurate?"
Every AI output is reviewed by the supervising attorney before it informs any decision or goes into any filing. Accuracy isn't guaranteed by the AI tool, it's guaranteed by the attorney who reviews it. That human review process is what professional responsibility requires, and what clients deserve.
Common Misconceptions About AI Disclosure
AI Myth vs. Reality
| Misconception | Reality |
| "Disclosure means admitting we did less work." | Disclosure is about transparency, not effort. Most clients understand that AI-assisted work still requires attorney judgment, review, and accountability. The quality of the output is what clients are paying for. |
| "One line in the engagement letter covers everything." | ABA Formal Opinion 512 is clear: boilerplate buried in an engagement letter is generally not sufficient for informed consent. Disclosure has to be specific enough for a client to understand what they're agreeing to. |
| "Disclosure will scare clients away." | The evidence points the other way. Sophisticated clients, especially corporate clients, increasingly expect AI governance from their outside firms. Transparency tends to build trust, and the firms that handle it proactively are set apart. |
| "This only matters for AI-generated documents." | Disclosure obligations can be triggered by AI use in research, document review, analysis, and billing. Firms that think narrowly about when AI disclosure applies may be underestimating their exposure. |
Why Transparency About AI Is a Competitive Advantage
Firms that have made AI disclosure a standard practice report a consistent outcome: clients appreciate it.
That shouldn't be surprising. Transparency is a trust signal. A firm that proactively tells a client how AI is being used, how their data is protected, and how billing works is a firm that's operating with confidence in its own practices. That confidence is visible, and clients notice it.
In competitive pitches and RFP responses, a clear AI governance narrative differentiates firms in a market where many competitors are still vague. Corporate legal departments are increasingly sophisticated buyers of legal services, and AI governance is now a factor in vendor selection in the same way data security and conflicts management have been for years.
See how firms using Filevine have operationalized AI-powered practice management while maintaining the oversight and documentation standards their clients expect.
There's also a longer-term dimension. As AI use becomes more widespread, and as clients become more sophisticated about how it works, the firms that established clear disclosure practices early will have an advantage. They'll have the templates, the training, the recordkeeping, and the client relationships built on transparency. The firms that treated disclosure as optional will be left catching up.
How Large Law Firms Can Get AI Disclosure Right
Disclosure isn't a one-time event. It's a practice. It’s repeatable, documented, and built into how the firm engages every client.
For large firms, that means three things working together:
- A clear firmwide policy that defines what AI use looks like and how it's governed.
- Standardized templates that give attorneys consistent language without requiring them to draft it themselves.
- Systems that make documentation and recordkeeping manageable at scale.
The ethics guidance is still evolving, and the jurisdiction-by-jurisdiction variation means firms need to monitor it actively. But the core of a good disclosure framework isn't complicated: tell clients what you're using, explain how their data is protected, describe the human review that guarantees the work, and make it easy for them to ask questions.
Firms that do this well don't treat disclosure as a compliance burden. They treat it as a reflection of how they run a good firm.
Frequently Asked Questions
Are law firms required to tell clients they use AI?
Yes, in many circumstances. Though "required" depends on context and jurisdiction.
ABA Formal Opinion 512 (2024) makes clear that AI disclosure may be required under Model Rules 1.4 (communication), 1.6 (confidentiality), and 1.5 (fees) when AI use is material to the representation.
Disclosure is most clearly required when AI processes confidential client information, meaningfully informs significant decisions, or affects billing. State bars and some courts have added their own requirements on top of the ABA framework, so firms should confirm the specific rules that apply in their jurisdictions.
Does AI use need to be written into the engagement letter?
Yes, but the engagement letter alone isn't always sufficient. ABA Formal Opinion 512 specifically cautions that boilerplate language buried in an engagement letter is generally not considered informed consent.
A clear, specific disclosure in the engagement letter is a good foundation, but for matters where AI use is significant, firms should supplement it with a more detailed conversation or a standalone AI use policy that clients can actually understand and evaluate.
Can a law firm bill clients for AI tools?
Yes, but it requires transparency. Model Rule 1.5 requires that fees be reasonable and that billing practices be disclosed to clients. Many firms treat routine AI use as overhead — absorbed into hourly rates the same way word processing software is. When AI delivers significant efficiencies, some firms pass through a cost or build it into a value-based fee structure.
Whatever approach the firm takes, clients should know before the bill arrives how AI use will be reflected in what they're charged.
Do all states have the same rules for disclosing AI use?
No. State bar guidance varies significantly. California, Florida, New Jersey, New York (including the NYC Bar's Formal Opinion 2024-5), and Pennsylvania have all issued guidance, and the rules and recommendations differ.
Some courts have gone further, requiring attorneys to certify AI use in certain filings. The landscape is still evolving, which means large firms with multi-jurisdiction practices need to monitor state-level developments actively and build disclosure frameworks flexible enough to adapt as requirements change.