Client Communication and AI: When Do You Have to Tell Clients You're Using It?
8/21/20263 min read
Client Communication and AI: When Do You Have to Tell Clients You're Using It?
"Do I have to tell my client I used AI?" is one of the most common questions attorneys ask right now — and the honest answer is: it depends, and the trend is toward more disclosure, not less. There's no bright-line rule requiring you to announce every use of a spell-checker-with-a-brain. But there are real situations where staying silent is the wrong call.
This is general information, not legal advice.
The duty behind the question: Rule 1.4
The relevant duty is Rule 1.4 (communication with clients). It requires you to keep clients reasonably informed and to give them enough information to make informed decisions about their representation. AI didn't create this duty; it created new situations where the duty bites.
The framing that's emerging in California is contextual, not categorical. The State Bar's proposed guidance points toward disclosure when AI use is significant enough to materially affect the scope, cost, manner, or decision-making of a representation. The routine, low-stakes use of an AI tool to tidy up formatting generally doesn't trigger a communication obligation. Using AI in a way that shapes strategy, changes what the client pays, or affects how the matter is handled is a different story.
Where the analysis actually lands
Think of it as a spectrum rather than a switch.
Disclosure generally not required
Low-risk, behind-the-scenes uses that don't affect the substance, cost, or direction of the representation — e.g., minor drafting assistance you fully review and verify.
Disclosure likely appropriate or advisable
AI use that materially affects strategy, work product the client relies on, or the client's decisions.
Situations that change the cost structure or how the matter is staffed and handled.
Where the client has asked, or has their own policies or expectations about AI use.
Where a court, not the client, requires it — some filings carry their own AI-disclosure requirements, which are separate from your Rule 1.4 duty to the client.
Consent, not just disclosure
Some contexts call for informed consent, not just a heads-up — particularly where confidentiality or a vendor's data handling is implicated. Telling the client is the floor; sometimes you need their agreement.
The contextual factors that push toward disclosure include the novelty and risk of the technology, the scope of the representation, and the client's own sophistication and expectations.
Practical ways to handle it
You don't need to renegotiate the relationship every time you open an AI tool. A few durable practices cover most situations:
Address it in the engagement letter. A clear statement that the firm may use AI tools, how it protects confidentiality, and how any related costs are handled sets expectations up front and reduces case-by-case guesswork.
Disclose proactively when it's material. If AI meaningfully shaped strategy or work product, say so in plain terms.
Honor client preferences. If a client asks about or restricts AI use, that governs — document it.
Keep billing honest. Don't bill for hours AI saved you as if the work were done manually; align your billing with the actual effort and any efficiencies gained.
Log your decisions. A short disclosure-decision record for matters where AI played a material role makes your reasoning demonstrable.
Why this is worth getting ahead of
Client expectations are shifting quickly, and so is the regulatory posture. Building disclosure into your engagement letters and your firm's AI policy now — rather than reacting to a client question later — is the low-friction way to stay both compliant and trusted. Transparency tends to protect the relationship as much as it protects you. (See our related pieces on building an AI use policy and attorney confidentiality with public AI tools.)
Frequently asked questions
Is there a rule requiring attorneys to disclose AI use to clients?
Not a blanket one. The duty flows from Rule 1.4's requirement to keep clients reasonably informed, and disclosure is expected where AI use materially affects the scope, cost, or decision-making of the representation.
Do I need to disclose using AI for legal research?
Often not, if the use is routine and you independently verify the output. But if it materially shapes the strategy or work product the client relies on, disclosure becomes appropriate.
Is court-ordered AI disclosure the same as telling my client?
No. Some courts impose their own AI-disclosure requirements for filings. Those are separate from — and in addition to — your Rule 1.4 duty to communicate with your client.
Does disclosure require client consent?
Not always, but sometimes. Where confidentiality or vendor data handling is implicated, informed consent — not just notice — may be the right standard.
This article is general information, not legal advice. Confirm the current State Bar guidance and any applicable rule before relying on specifics.
Sources
California Rules of Professional Conduct 1.4 (communication with clients), 1.6 (confidentiality)
State Bar of California, Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law (2023; updated 2026), and proposed amendment adding a Rule 1.4 comment on AI disclosure
ABA Formal Opinion 512 (2024) — addressing AI disclosure obligations to clients