.png){kind=small}
Can Legal Marketing Use AI?
Key Takeaways
Short Answer: Yes. The ABA, the Florida Bar, and Google all allow AI in legal marketing, but each one judges what you publish, not the tool that drafted it.
Ethics Strings Attached: ABA Formal Opinion 512 and Florida Bar Opinion 24-1 tie AI use to competence, confidentiality, supervision, and the advertising rules that already govern your firm.
The SEO Trap: Google doesn't penalize content for being AI-generated, but pages mass-produced to game rankings fall under its scaled content abuse policy.
Why Generic Tools Backfire: Blank-slate chatbots default to clichés, implied guarantees, and invented facts, which hurts rankings and creates risk under Rule 7.1.
What Actually Works: A custom AI persona configured with your voice, your jurisdiction's rules, and a privacy-first platform, with a human approving every draft before it goes live.
AI Can Broaden Your Horizon and Reach
Yes, legal marketing can use AI, and if your hesitation comes from seeing an AI draft that sounded like a stranger guessing at your practice area, you're worried about the right problem in the wrong place. The tool isn't what goes wrong. The configuration is, along with the habit of publishing whatever falls out of a blank chatbot window.
We're Dailey Curation LLC, and we build custom AI personas for law firms that want AI's speed without its usual mess. Our founder has spent more than seven years managing content teams and writing, editing, and designing web pages and infographics for law firms, and she has edited and curated thousands of law firm pages and blog posts along the way. She also started out as an AI skeptic, which is exactly why everything we build is designed around human control rather than blind trust.
Below, we walk through what the bar actually says, what Google actually penalizes, why generic AI fails both tests, and how a properly configured persona fixes the problem. If your marketing feels stuck somewhere between "too slow" and "too risky," tell us where your marketing is stuck.
What the Bar Actually Says About AI-Assisted Marketing
Plenty of attorneys still assume AI marketing is either quietly forbidden or a total free-for-all. Neither is true, and both assumptions cost firms. One keeps them paying for slow, expensive content, while the other walks them toward a grievance file.
Opinion 512 Treats AI Like Any Other Tool You Answer For
On July 29, 2024, the ABA's Standing Committee on Ethics and Professional Responsibility released its first formal opinion on generative AI, Formal Opinion 512. The committee's core point is that the existing rules on competence, informed consent, confidentiality, and fees principally apply. In other words, nothing new got banned, and nothing old got waived.
For marketing, competence is the rule with teeth. You're expected to understand what a tool can and can't do before you rely on it, and generative AI is famously confident when it's wrong. A lawyer who doesn't know that AI invents sources isn't using it competently, no matter how polished the output looks.
If that sounds abstract, remember Mata v. Avianca. In June 2023, a federal judge in Manhattan fined two lawyers and their firm $5,000 after they submitted a brief citing cases ChatGPT had fabricated. That was a court filing rather than a blog post, but the root failure is identical: nobody checked.
Florida Put the Advertising Rules in Writing
The Florida Bar got specific even earlier, in January 2024. Florida Ethics Opinion 24-1 says generative AI chatbots that communicate with clients or third parties must comply with lawyer advertising rules and disclose that they're an AI program, not a lawyer or firm employee. The opinion also puts the lawyer on the hook if an intake chatbot misleads prospective clients or communicates in an inappropriately intrusive or coercive way.
Florida's opinion also applies its supervision rule for nonlawyer assistants to generative AI. That's the most useful framing we've seen for marketing teams. Treat AI like a talented, overconfident new hire whose work never leaves the building unreviewed.
Even if you don't practice in Florida, the opinion is worth reading as a model. It takes rules every state already has and simply points them at a new kind of assistant. Your jurisdiction may phrase things differently, but the logic travels.
Rule 7.1 Doesn't Care Who Drafted the Sentence
Every state has some version of the rule against false or misleading communications about a lawyer's services. Colorado's Rule 7.1, which tracks the ABA model language, makes clear that even truthful statements are prohibited when they're misleading. California's comments go a step further, stating that an express guarantee of results counts as a false or misleading communication.
Here's why that matters for AI. General-purpose models learn from oceans of marketing copy, and marketing copy loves a promise. Ask one for a personal injury page and you'll regularly get some flavor of "we'll get you the compensation you deserve," which is exactly the kind of sentence an ethics reviewer circles in red.
The table below translates the main rules into plain marketing terms.
Rule or Opinion | What It Requires | What It Means for Your Marketing |
ABA Formal Opinion 512 | Competence, confidentiality, client communication, and reasonable fees when using generative AI | Know your tool's limits and verify every factual claim before publishing |
Florida Bar Opinion 24-1 | Advertising compliance and AI disclosure for client-facing chatbots | Intake bots must identify as AI and can't say anything you couldn't say yourself |
Rule 7.1 (in every state, in some form) | No false or misleading communications, including misleading truths | No implied guarantees, inflated results, or unverifiable superlatives in AI drafts |
Supervision rules for nonlawyer assistance | Reasonable oversight of work done on the lawyer's behalf | A named human reviews and approves every AI-assisted piece |
None of this says "don't use AI." It says the person whose name is on the bar card owns every word. If you want a compliance-first setup mapped to your state's rules, reserve a free compliance walkthrough with us.
Can Legal Marketing Use AI and Still Rank on Google?
Here's the uncomfortable truth we share with every firm: almost nobody reads your blog closely. Prospective clients skim a headline and a paragraph on their phones, then call or leave. Search engines, on the other hand, read every word, every heading, and every link, which makes them your most attentive audience by a wide margin.
That's why the SEO question matters more than the style question. If a search engine decides your content is filler, it doesn't matter how nicely it reads to the managing partner. We dig into that dynamic in why no one reads your blog, but the short version is that your content has to earn its place with machines before it ever reaches a human.
So can legal marketing use AI without sabotaging search visibility? It can, as long as you understand what search engines are actually measuring. Spoiler: it isn't whether a human typed every keystroke.
Google's Position Is Clearer Than the Rumors
Google addressed this directly in February 2023. Its guidance on AI-generated content says the company focuses on content quality rather than how content is produced, and that using AI primarily to manipulate search rankings violates its spam policies. The same guidance notes that AI content gets no special ranking advantage simply for being AI.
In March 2024, Google sharpened that position with a new scaled content abuse policy. It targets large numbers of pages generated mainly to manipulate rankings instead of helping users, whether those pages come from automation, human effort, or a mix of both. A firm pumping out fifty near-identical "car accident lawyer in [city]" posts from a chatbot is doing almost exactly what that policy describes.
The takeaway is reassuring and demanding at the same time. AI isn't the enemy of your rankings. Thin, interchangeable, purpose-free content is, and AI just makes it much easier to produce at volume.
Why Generic AI Copy Loses the Ranking Game
Google's guidance leans on E-E-A-T: experience, expertise, authoritativeness, and trustworthiness. It also encourages creators to think about who made the content, how it was made, and why it exists. Generic AI copy struggles with all three, because it has no experience, borrows its expertise from the average of the internet, and usually exists to fill a content calendar.
The deeper problem is sameness. When thousands of firms ask the same tools similar questions, they get similar answers, and search engines have little reason to rank one generic page over another. The pages that win contain something a model couldn't have guessed on its own, like a local court quirk, a firm's actual intake process, or a genuine point of view.
That's the gap a configured persona fills. It doesn't make AI smarter about the law. It makes AI start from your firm's knowledge, voice, and rules instead of the internet's average.
Where Off-the-Shelf Chatbots Go Wrong for Law Firms
A blank chatbot has no idea who you are. It doesn't know your jurisdiction, your practice areas, your tone, or which claims your bar would object to. So it fills those gaps with the most statistically likely answer, which is almost always the most generic one.
We see the same failure patterns across enough drafts that we can nearly predict them. Firms that try a general-purpose chatbot cold usually end up with some combination of the following problems:
Recycled clichés like "aggressive representation" and "we fight for you," which say nothing a competitor's page doesn't already say.
Implied guarantees about outcomes or compensation, the exact language Rule 7.1 exists to stop.
Invented statistics, studies, or statutes that sound authoritative and don't exist anywhere.
Laws from the wrong state, or national generalizations dressed up as local rules.
A voice that matches nobody at the firm, which clients notice the moment they call and hear a real person.
Every one of these is fixable by a careful editor. The catch is that fixing them every single time eats the hours AI was supposed to save. If a partner spends forty minutes cleaning up a ten-minute draft, the math stopped working a while ago.
The Prompting Treadmill
Some firms try to solve this with longer prompts. They paste in a paragraph about tone, a list of banned phrases, and a reminder to cite sources, over and over. It works for a draft or two, then someone forgets a line and the clichés creep right back in.
That's not a discipline problem. It's a design problem, and it's the reason ad hoc prompting doesn't hold up for firms with real compliance obligations. The instructions need to live somewhere permanent, not in whoever happens to remember them on a busy Tuesday.
There's also a consistency cost. Three people prompting three different ways produce three different firm voices, and readers who land on more than one page can feel the seams. Search engines notice uneven quality, too, even if no single page is obviously bad.
Confidentiality Is a Marketing Issue, Not Just a Litigation One
Attorneys tend to think about confidentiality in terms of case files and briefs. Marketing feels safer by comparison. Yet some of the strongest marketing material a firm has, like a case result, a client story, or a trend spotted during intake, comes straight out of confidential information.
Pasting those details into a consumer chatbot raises the same questions Opinion 512 raises for any client data. Where does the information go, who can access it, and is it used to train future models? Lawyers clearly feel the tension, since data privacy and security ranked among the top three AI concerns in the ABA's 2024 Legal Technology Survey Report, cited by 47% of respondents according to LawNext's coverage of the survey.
Adoption is climbing anyway. That same report found 30% of responding attorneys using AI-based tools in 2024, up from 11% the year before, with accuracy topping the list of concerns at 75%. Firms are clearly deciding to use AI, so the real question is whether they're setting it up to protect what matters.
Choosing Where Your Drafting Happens
Platform choice is part of the answer. GreenPT, the sustainable, privacy-first AI platform we partner with, keeps data inside the European Union and states in its published privacy commitments that user conversations are never used to train its models. It also runs on renewable energy, which matters to firms that care about the footprint of their tools as much as the output.
Platform choice isn't the whole answer, though. Even on a private platform, the safest marketing workflow keeps identifiable client details out of drafts entirely. That rule belongs inside the AI's configuration, not on a sticky note next to someone's monitor.
This is where a persona earns its keep again. We build personas that anonymize results by default, push back when a request includes identifying details, and remind users when a draft edges toward confidential territory. The guardrail works the same way every time, regardless of who's typing.
What a Custom AI Persona Changes About the Workflow
A custom AI persona is a configured assistant, built as a GPT, a Gemini Gem, a Claude Project, or a GreenPT setup, that starts every conversation already knowing your firm. Instead of retyping the same instructions every time, the rules live inside the persona. Your team opens it and starts writing.
We build each persona around the specific ways generic AI fails law firms. In practice, that means configuring it with these five layers:
Your firm's voice, drawn from your strongest existing pages, so drafts sound like your attorneys instead of a template.
Your jurisdiction's advertising rules, including banned phrasing like outcome guarantees and unverifiable superlatives.
Sourcing rules that require named, verifiable sources and flag any statistic the AI can't support.
SEO structure built in, including heading hierarchy, keyword placement, and the internal linking habits search engines reward.
Confidentiality guardrails that keep identifying client details out of drafts.
None of that replaces the attorney's final read. It shrinks it. A reviewer checking a draft that already follows the rules is approving work, not rescuing it.
Here's how the two approaches compare on the tasks that matter most.
Task | Generic Chatbot | Custom AI Persona |
Voice | Defaults to stock legal marketing phrasing | Configured with your firm's own pages and preferences |
Advertising compliance | Relies on someone remembering to ask | Bar rules and banned phrases built into its instructions |
Sources and statistics | May invent figures that look real | Required to name sources and flag anything unverified |
SEO structure | Varies from prompt to prompt | Consistent headings, keyword use, and linking patterns |
Confidentiality | Depends on what users paste in | Guardrails that discourage identifiable client details |
Human review | Heavy rewriting on every draft | Lighter, faster approval of rule-following drafts |
If you've considered outsourcing instead, remember that handing content to an outside agency doesn't hand off your responsibility for it, a point we cover in our guide to outsourcing content creation. A persona keeps drafting in-house, where your attorneys can see and approve everything. You keep the control and lose the bottleneck.
Human Oversight Isn't Optional
We'll say it plainly because the rules said it first: AI-assisted content still needs a human who answers for it. Opinion 512, Florida's supervision framing, and Google's emphasis on trustworthiness all land in the same place. Someone with judgment reads it before the world does.
Our personas are designed for that handoff. They draft, structure, and flag, while your team verifies, edits, and publishes. That division of labor is what makes AI safe enough for small firms to use every week, not just experiment with once.
Yes, AI Belongs in Legal Marketing, on a Short Leash
So, can legal marketing use AI? Yes, and for small firms with tight budgets, it may be the most practical way to keep up with the kind of content search engines actually reward. The bar hasn't banned it, Google hasn't banned it, and both have told you exactly what they care about: accuracy, honesty, confidentiality, and real expertise.
We didn't arrive at that view easily. Our founder spent years managing content teams and editing thousands of law firm pages and blog posts, and she started out deeply skeptical of AI. What changed her mind wasn't the hype getting louder, but seeing what AI could do once a human configured it carefully and kept the final say over everything it published.
That's the entire premise of our custom AI personas. We build them to sound like your firm, follow your jurisdiction's rules, cite real sources, respect client confidentiality, and structure every page for the search engines that are, honestly, your most devoted readers. Drop us a line about your firm's content goals and we'll tell you what a persona could take off your plate.
If you'd rather see the setup before committing to anything, we're happy to show you how it works. Claim a free consultation slot to map out a persona for your practice. And if you want to know who's behind the work first, get to know our team and process.
Sources
American Bar Association, ABA ethics guidance on AI tools, announcing Formal Opinion 512 (July 29, 2024).
Bloomberg Law, ChatGPT brief sanctions in Mata v. Avianca (June 22, 2023).
The Florida Bar, Florida Ethics Opinion 24-1 on lawyers' use of generative AI (January 19, 2024).
Colorado Bar Association, Colorado Rule 7.1 text, Communications Concerning a Lawyer's Services.
State Bar of California, California Rules of Professional Conduct, Chapter 7, Information About Legal Services.
Google Search Central, Google's AI-generated content guidance (February 8, 2023).
Google Search Central, March 2024 spam policies update, introducing the scaled content abuse policy (March 5, 2024).
LawNext, coverage of the ABA tech survey, reporting results of the ABA's 2024 Legal Technology Survey Report (March 2025).
GreenPT, published privacy commitments page on EU data hosting and model training.
Comments