Picture this: a firm used generative AI to draft a client-facing brief, and because it lacked proper AI disclaimers for law firms in its engagement letter, a bar grievance landed on the managing partner’s desk months later. The firm billed the full hourly rate as if an associate wrote every word and never mentioned AI anywhere in the client file. The issue wasn’t the AI itself. The issue was that the client was never told.
That scenario is no longer hypothetical. In 2026, bar opinions, court standing orders, and state-level advertising rules have made AI disclosure language a required part of how law firms operate in specific, well-defined circumstances, not an optional courtesy. At Thrive Business Marketing, helping law firms get this language right is a core part of how we build ethical AI marketing frameworks for our clients from day one. Getting the disclosure right from the start protects the firm and builds the kind of client trust that turns one matter into a long-term relationship.
This guide tells you exactly when disclosure is legally and ethically required, what elements your disclosure needs to cover, and hands you eight sample clauses you can copy and adapt today for engagement letters, court filings, and marketing content.
When bar rules actually require AI disclaimers for law firms (and when they don’t)
ABA Formal Opinion 512, issued in July 2024 and reaffirmed in 2025 guidance, sets the national baseline. It does not require blanket disclosure. Using AI to speed up legal research is not automatically a disclosure event, any more than using a word processor is. The opinion identifies three specific triggers that make disclosure mandatory: materiality, novelty, and billing impact.
The billing scenario is the clearest. If AI reduces the time required to complete a task and the client is still billed at the pre-AI hourly rate, disclosure is mandatory. The same applies if the firm passes AI tool costs directly to the client. Opinion 512 also requires informed consent whenever confidential client information is input into a third-party or self-learning AI system, and that consent must be fact-specific, not buried in boilerplate.
State bars frequently go further than the ABA baseline. California requires affirmative disclosure whenever generative AI drafts, summarizes, or analyzes documents for a client or court filing, unless a prior written agreement says otherwise. New York City uses a “substantial use” threshold. Florida applies a consent-to-material-change standard. Kentucky is the most permissive, requiring disclosure only when work is outsourced or AI tool costs are billed to the client. Your state bar’s rules govern your conduct, and they often exceed ABA guidance.
Across jurisdictions, the practical standard most attorneys can apply is this: if the client would reasonably want to know that AI shaped the work product, disclose it. That test covers edge cases that don’t fit neatly into materiality analysis, including AI-drafted intake summaries, demand letters, and client communications.
What courts are demanding in AI-related filings
Over 30 federal district courts now have some form of standing order or local rule governing AI use in filings. At least 15 of those orders require attorneys to name the specific tool used, not just reference “AI software” generically. The language in these orders varies, but the structure is consistent.
Consider the range of requirements already in effect. Judge Brantley Starr in the Northern District of Texas requires attorneys to certify that no portion of a filing was AI-drafted, or to identify AI-drafted sections and confirm a human reviewed them. Judge Evelyn Padin in the District of New Jersey requires naming the specific tool (e.g., “ChatGPT-4” or “Claude”), identifying which parts of the filing were affected, and certifying human review. Judge Dana Sabraw in the Southern District of California requires disclosure of “the name, version, and provider of any generative AI tool used” in briefs, motions, or memoranda. The District of Kansas went further in January 2026 with a court-wide standing order, the first uniform AI disclosure rule covering every case in that district.
Judge Easterbrook of the 7th Circuit draws a useful distinction: disclosure is required when AI is used for substantive legal analysis or research, but not for grammar, formatting, or other ministerial tasks. That distinction matters when you’re drafting internal firm policies on what triggers a certification obligation.
Most recently issued court orders include a human verification requirement. Attorneys must certify that a licensed attorney reviewed and verified all citations and legal assertions, regardless of how the document was initially drafted, though the exact certification language varies by court. This is the courts’ mechanism for allocating responsibility: the attorney owns the filing, not the algorithm. Your internal AI policy and client-facing disclosures should use the same framing.
State courts are building their own patchwork. New York’s Uniform Court System guidelines, effective June 1, 2026, require citation verification but stop short of mandating full disclosure. Florida’s Miami-Dade and Broward counties require a specific filing statement certifying independent review of all factual assertions and legal authority. California has a proposed rule pending in several superior courts. The only safe approach is to check the specific judge’s standing order and local rules before every filing.
The four elements every effective AI disclaimer must address
Vague language creates more exposure than no language at all. A clause that says “we may use technology including artificial intelligence” is not informed consent under ABA Opinion 512, and it won’t satisfy the courts or state bar guidelines already in force. Effective AI disclaimers for law firms address four specific elements.
Confidentiality and data protection must be explicit. The strongest clauses specify that the firm uses only enterprise-grade platforms that contractually prohibit using client data for model training, and that privileged communications are not input into AI tools without prior written consent. Naming that commitment directly builds more client trust than generic assurances about “data security.”
Accuracy limits require honest acknowledgment. Phrases like “all AI-generated work product is reviewed, verified, and approved by licensed attorneys before being relied upon or delivered to clients” reflect the industry standard in 2026, consistent with guidance from sources such as the State Bar of Texas AI Toolkit and ABA model language. Vague language about “AI assistance” without a human review commitment actually increases malpractice exposure by implying reliance on unreviewed output.
Vendor transparency and liability allocation are increasingly required by specific courts and represent good practice everywhere. Naming the specific AI tool promotes transparency and may assist internal risk allocation if a system fails or causes a data breach, though firms should consult their malpractice carrier and counsel about specific liability implications. The disclaimer should state explicitly that attorneys, not algorithms, have final authority over all work product.
Client rights round out an effective disclosure. Clients should know they can ask which tools were used on their matter and, in many jurisdictions, request limits on AI use for specific work. Including that right in the engagement letter is both ethically sound and a practical differentiator for firms competing on client service.
Eight sample AI disclosure clauses ready for your firm
These clauses are adapted from published bar association guidance, including the Texas Bar’s AI Toolkit and templates aligned with ABA Model Rules. Each clause should be adapted to your jurisdiction and specific AI tools before use.
Engagement letter AI language (clauses 1, 4)
Clause 1: Basic AI use disclosure (modeled on the Texas Bar’s simple version; for firms using AI in research and drafting with full attorney review):
“The Firm may use reputable artificial intelligence tools to support legal research, drafting, and administrative tasks. These tools are used only under attorney supervision and do not replace professional legal judgment. All AI-generated work product is reviewed and approved by a licensed attorney before delivery to you.”
Clause 2: Enhanced disclosure with data protection (for firms using third-party AI tools that process client data):
“The Firm may use AI tools to assist with legal research, document review, and drafting. The Firm will not input identifiable or privileged client information into any AI system that stores or uses data for future training, unless the system operates in a private enterprise environment with contractual prohibitions on training-data use. All AI use complies with applicable ethical rules, confidentiality obligations, and data protection requirements.”
Clause 3: Client opt-out right clause (adapted from industry engagement letter templates aligned with ABA guidance on client autonomy):
“Client may, at any time, request that the Firm limit or exclude the use of AI tools in connection with specific matters or categories of work. Such requests should be made in writing to the supervising attorney. Limitations on AI tool use may affect the timeline and cost of deliverables.”
Clause 4: Billing and fee disclosure clause (for firms where AI reduces billable hours or tool costs are passed to the client, fee treatment should be tailored to your jurisdiction and client agreement):
“Where AI tools materially reduce the time required to complete a task, the Firm will disclose the impact on billing and may reflect that reduction in the fees charged. Any separately incurred AI platform subscription costs passed to the client will be identified as a line item on invoices and will not exceed the firm’s actual cost.”
Client-facing notices and website disclaimers (clauses 5, 6)
Clause 5: Website or chatbot disclaimer (prevents implied attorney-client relationship formation; recommended when AI chatbots handle initial inquiries):
“This website uses automated tools, including AI-assisted chat, to respond to initial inquiries. Use of this website or its chat feature does not create an attorney-client relationship. No information submitted through this site should include confidential or privileged details until a formal engagement letter has been signed.”
Clause 6: AI-assisted marketing content notice (for blog posts, social content, or email campaigns drafted with AI assistance):
“Some content on this website is drafted with the assistance of AI writing tools and reviewed by licensed attorneys for accuracy and compliance with applicable bar advertising rules. Nothing on this site constitutes legal advice or creates an attorney-client relationship.”
Court filing certifications (clauses 7, 8)
Clause 7: General court filing AI certification (human verification confirmed, tool named, scope of use described):
“Counsel certifies that [Tool Name, e.g., Claude 3.5] was used in the preparation of this filing for [describe scope, e.g., initial research and draft organization]. All legal citations, factual assertions, and legal analysis contained in this filing have been independently reviewed and verified by the undersigned attorney.”
Clause 8: Substantive AI use certification (for jurisdictions following Judge Easterbrook’s 7th Circuit standard):
“Counsel discloses that generative AI was used for substantive legal research and drafting in this filing. AI was not used for grammar correction, formatting, or other ministerial tasks. The undersigned attorney has reviewed and takes full professional responsibility for all content, including all cited authorities, which have been verified using primary legal research databases.”
AI disclaimers for law firms in marketing content: where the gap usually lives
Most attorneys think about AI disclosure in the context of court filings and engagement letters. Far fewer have considered their blog posts, social ads, practice area pages, or attorney bio updates written with AI assistance. That gap is where bar advertising violations happen quietly, often without anyone at the firm realizing they have exposure.
Under ABA Model Rule 7.1, all attorney advertising must be truthful and non-misleading. Under Rule 5.3, attorneys are responsible for supervising anyone producing content under the firm’s name, including marketing vendors using AI. If AI-generated marketing content misrepresents a firm’s practice areas, fabricates case outcomes, or includes inaccurate legal claims, the firm faces a bar advertising violation, not just a content quality problem. Multiple state bars have indicated that AI-generated advertising materials require the same attorney supervision and accuracy review as any other marketing material.
This is the specific gap that Thrive Business Marketing is built to close. Thrive works exclusively with law firms and understands both the marketing opportunity and the compliance obligations that come with bar advertising rules. When Thrive produces AI-assisted content for a law firm, every piece goes through review by legal marketing specialists familiar with state advertising rules before it goes live. That’s not just quality control, it’s the governance layer that protects the firm’s license while building its online presence.
Thrive also helps law firms develop the AI use policies and disclosure standards that govern how marketing content is produced internally. For firms that want compliance built in from day one, not added as an afterthought after a grievance, that kind of specialized partnership matters more now than it did two years ago.
Three things to do after reading this article
Start by mapping your current AI use against the ABA Opinion 512 disclosure triggers and your state bar’s specific rules. If you’re in California, New York, or Florida, your state rules almost certainly go beyond the ABA baseline, and your engagement letters need to reflect that.
Next, update your engagement letter. If your firm uses any third-party AI tool that processes client information, Clause 2 above should be your starting point. Data protection language is the highest-priority addition for most firms in 2026 because it addresses the ABA’s informed consent requirement for client data directly.
Finally, audit your marketing content. Pull the last ten pieces of content published on your firm’s website or social channels and ask two questions: Was any of it AI-generated or AI-assisted? Did a licensed attorney review it for accuracy before publication? If the answer to the first is yes and the second is no, you have a disclosure and supervision gap to close.
AI disclaimers for law firms are now a baseline expectation in 2026, from bar associations, courts, and clients who are increasingly asking the question directly. Thrive Business Marketing helps law firms build that compliance infrastructure into their marketing strategy from the start, so growth never comes at the expense of a firm’s ethical standing.