Law firms are being told two opposite things about AI content: that it will flood the web with pages for free, and that Google and the bar will punish anyone who uses it. Neither is right. The bar opinions issued so far allow lawyers to use generative AI, with supervision and checking. Google says it judges content on quality, not on how it was made, while treating mass-produced pages made to rank as spam. This guide sets out what each actually says, with links, and the review steps that keep AI-assisted content inside both. It sits alongside our law firm SEO guide, which covers what content to write in the first place.
What ABA Formal Opinion 512 says
On July 29, 2024, the ABA’s Standing Committee on Ethics and Professional Responsibility issued Formal Opinion 512, its first opinion on generative AI (ABA). It doesn’t create new rules. It applies the existing Model Rules to a new tool. The points that matter most for marketing content, as summarized in NCBE’s The Bar Examiner (NCBE):
- Competence (Rule 1.1). A lawyer using AI needs a reasonable understanding of what the tool can and can’t do. AI can’t substitute for a lawyer’s own competent work.
- Confidentiality (Rule 1.6). For self-learning tools, informed client consent is required before entering information relating to a representation. Boilerplate in an engagement letter isn’t informed consent.
- Supervision (Rules 5.1 and 5.3). Managing lawyers should set clear firm policies on permissible AI use.
- Fees (Rule 1.5). Lawyers billing hourly bill only their actual time, and shouldn’t charge clients for their own inexperience with a tool.
Two things follow for a firm’s website. First, a person with legal judgment has to check what an AI draft says, because the firm is responsible for it. Second, client facts don’t go into a tool to make a “case study” without the attorney’s decision and, where required, the client’s informed consent.
Formal opinions interpret the ABA Model Rules. Your state’s rules and its bar’s opinions are what bind you.
What state bars have said
Several state bars have issued their own AI guidance. The table covers ones we checked on their official sites. Each says broadly the same thing as the ABA. Florida’s goes furthest on advertising.
State bar AI opinions, checked October 24, 2026
Six opinions, one message: use it competently, protect clients, check everything
| Body | Opinion | Date | What it says that matters for marketing |
|---|---|---|---|
| ABA | Formal Opinion 512 | July 29, 2024 | Competence, confidentiality, supervision and fees apply to AI; boilerplate consent isn’t enough for self-learning tools |
| Florida Bar | Ethics Opinion 24-1 | January 19, 2024 | AI use is ethical only to the extent compliance can be ensured; advertising rules apply; a chatbot must tell prospective clients they’re talking to an AI program; no claims that the firm’s AI is superior unless objectively verifiable |
| State Bar of Texas, Professional Ethics Committee | Opinion 705 | February 2025 | Lawyers can’t blindly rely on AI output; must verify it; must understand the tool’s risks; treats AI like a nonlawyer assistant to supervise |
| State Bar of California | Practical Guidance for the Use of Generative AI | November 16, 2023; revisions approved May 14, 2026 | Guiding principles applying the Rules of Professional Conduct and the State Bar Act to AI use |
| D.C. Bar | Ethics Opinion 388 | April 2024 | Understand how the tool works; vet whether the provider stores inputs; check output for accuracy, especially citations |
| North Carolina State Bar | 2024 Formal Ethics Opinion 1 | November 1, 2024 | AI may be used competently, securely and with supervision; avoid entering client-specific information into public tools; the lawyer stays responsible for output |
The Florida opinion is the one to read if you plan to put AI in front of the public. It warns lawyers to be careful when using a generative AI chatbot for advertising and intake, and it applies the state’s rule against deceptive or inherently misleading advertising (Florida Bar). If your site has a chatbot, the opinion says prospective clients must be told they are communicating with an AI program, and that it is not a lawyer or law firm employee.
The advertising rules don’t change
An AI draft becomes lawyer advertising the moment it’s published. Every rule in your state applies to it exactly as it would to a page a partner wrote: no misleading statements, care with past results and testimonials, any required disclaimers, and in some states, filing. AI drafts tend to break these rules in predictable ways:
- Inventing results or credentials. A model asked to “make it persuasive” may add a verdict, an award or a years-of-experience figure that isn’t true.
- Superlatives. “The leading”, “the most experienced”, “the most trusted”: comparisons many states restrict unless they can be proven.
- Wrong law. Deadlines, damages caps and statute numbers that are out of date or from another state.
- Implied promises of outcomes. “We’ll get you the compensation you deserve.”
Our lawyer advertising rules guide covers the rules state by state. Check every AI draft against them before it reaches your responsible attorney.
What Google says about AI content
Google’s position has been consistent since February 2023. Its ranking systems “aim to reward original, high-quality content” that shows experience, expertise, authoritativeness and trustworthiness, however it is produced (Google). And appropriate use of AI is not against its guidelines. What is: using automation, including AI, to generate content with the primary purpose of manipulating search rankings, which breaks Google’s spam policies (Google).
Three Google documents give the detail.
1. Scaled content abuse. Google’s spam policies define it as many pages generated for the primary purpose of manipulating search rankings, and give as an example using generative AI tools to generate many pages without adding value for users (Google). A firm that publishes a hundred near-identical “car accident lawyer in [town]” pages, written by AI, fits the description.
2. Using generative AI on your site. Google’s guidance says it is critical to fact-check and review all AI-generated content for accuracy before publishing, because models can produce hallucinations. It applies the same care to titles, descriptions, structured data and image alt text, and recommends telling readers how automation was used where that’s useful (Google).
3. Helpful, people-first content. Google asks publishers to think about “Who, How, and Why”: who created the content, with bylines where expected; how it was made, including whether automation was used; and why, which should be to help people rather than to attract search visits (Google). Among the warning signs it lists: “using extensive automation to produce content on many topics”. It also says its systems give more weight to trust signals on “Your Money or Your Life” topics, those that could significantly affect health, financial stability or safety (Google). Most legal topics fit that description.
Put together: AI isn’t the problem. Volume without value is, and on legal topics, accuracy is held to a higher standard.
A review process that satisfies both
Free checklist
Seven steps from AI draft to published page
| # | Step | Owner | Record kept |
|---|---|---|---|
| 1 | Choose a question real clients ask, from intake notes or search data | Editor | The question and where it came from |
| 2 | Collect primary sources first: statutes, court rules, agency pages | Writer | Source list with dates checked |
| 3 | Draft, with AI help allowed for structure and first wording, in English only | Writer | Draft version |
| 4 | Check every sentence of fact against a source; delete what can’t be sourced | Second reviewer | Marked-up draft |
| 5 | Check the state’s advertising rules: results, comparisons, disclaimers, filing | Reviewer, with the state rule list | Checklist |
| 6 | Attorney approval of the final text | The firm’s responsible attorney | Dated approval |
| 7 | Publish with a review date and sources; schedule the next review | Editor | Page history |
Two rules sit around the checklist. No client information goes into a public AI tool without your responsible attorney’s decision. And one page per real question. If two pages would say the same thing with a different town name, write one.
Questions to ask a vendor that uses AI
If an agency or freelancer writes for your firm, ask these before you sign, and get the answers in writing:
- Which AI tools do you use, and for what? Research, outlines and first drafts are different from publishing output unedited.
- What goes into the tools? Nothing about your clients or matters should, without your attorney’s decision.
- Who checks the facts? Ask for the name of the role, not the person, and what source list comes with each page.
- How many pages a month, and why that number? A high fixed page count with town names swapped in is the pattern Google calls scaled content abuse.
- Who checks the advertising rules? Ask which state’s rules they check against, and how.
- Who writes the Spanish? Ask whether a native speaker writes it and a second person reviews it.
- What happens when a page is wrong? There should be a correction process with a date on the page.
A vendor with good answers will have them ready. A vague answer to question 3 is the one to worry about most.
Disclosure and your editorial policy
Google suggests disclosing AI use where readers might reasonably wonder how a page was made (Google). The simplest way for a law firm is an editorial policy page that says how content is made, who reviews it, and how errors are corrected, linked from each article. Ours is our editorial policy: software, including AI tools, may help with research or a first draft in English; every sentence and source is checked by a person; and no Spanish we publish is drafted by AI.
Chatbots are different. If AI talks to prospective clients on your site, follow your state’s rules on identifying it. Florida’s opinion requires telling prospective clients they’re communicating with an AI program (Florida Bar).
Why we don’t AI-draft Spanish
Spanish-speaking clients deserve pages written for them, not translated at them. Legal Spanish has terms that differ by country of origin and that machine output often gets wrong or renders too literally, and the advertising rules apply to a Spanish page exactly as to an English one. So our Spanish is written by native speakers and checked by a second reviewer. Our Spanish-language legal marketing guide covers the rules on non-English advertising.
What this means for your firm
- Write a one-page AI policy for marketing content: which tools, what may never go into them, who reviews and who approves.
- Check every fact against a primary source, and delete what you can’t source.
- Run every draft past your state’s advertising rules before attorney approval.
- Publish fewer, better pages. One page per real question; no near-duplicate town pages.
- Publish an editorial policy that says how pages are made and corrected.
Probative is AI-first SEO and ads for law firms: AI-first means firms get named in AI answers because of the work, not that AI writes your pages unchecked. Our legal content service follows the steps above, with a state-rule check and your attorney’s approval on every piece. For how AI search engines choose what to cite, see our AI search guide.
Marketing information, not legal advice. Ethics opinions are advisory and change; check your own state’s rules and opinions, and ask your bar’s ethics hotline or your own counsel.