Should a law firm use AI to write its blog?

A law firm can use AI to help draft its blog, but not to supply the substance or to skip attorney review. ABA Formal Opinion 512 says lawyers using generative AI must consider duties including competence, confidentiality, communication, and supervision, and it warns that output can be plausible and wrong. The more useful question is what the AI drafts from: the firm's approved knowledge, or nothing in particular.

Most firm owners have seen the result of the second option. A post appears in minutes, reads smoothly, and could have come from any firm in any state. It may also contain a statement of law nobody checked.

This is a content-planning question for firm owners and marketing leads, not legal advice. Opinion 512 is based on the ABA Model Rules, and your state's rules and ethics opinions may say more.

What ABA Formal Opinion 512 says

The ABA issued Formal Opinion 512 on July 29, 2024. It says lawyers using generative AI tools "must fully consider their applicable ethical obligations, including their duties to provide competent legal representation, to protect client information, to communicate with clients, to supervise their employees and agents," along with duties on meritorious claims, candor toward the tribunal, and reasonable fees.

On accuracy, the opinion warns that some tools are "prone to 'hallucinations,' providing ostensibly plausible responses that have no basis in fact or reality." On supervision, it says "Managerial lawyers must establish clear policies regarding the law firm's permissible use of GAI," and that supervisory concepts "also apply to GAI providers and tools."

Two limits are worth stating plainly. The opinion focuses on legal work rather than marketing, though a footnote flags Model Rule 7.1 as a possible additional issue. And it calls the tools "a rapidly moving target," with a footnote tying its conclusion on client consent to "the risks and capabilities of GAI tools as of the publication of this opinion." Read it as a floor for your firm's AI policy, then check your own state's guidance.

For a marketing team, the practical consequence is simple. Someone at the firm with authority needs to approve which tools are used, what may go into them, and who reviews the output before it publishes.

What Google says about AI-written pages

Google does not prohibit AI-assisted content. Its guidance says "Generative AI can be particularly useful when researching a topic, and to add structure to original content." It also warns that "using generative AI tools or other similar tools to generate many pages without adding value for users may violate Google's spam policy on scaled content abuse," and says "It is critical to manually factcheck and review all AI-generated content for accuracy and trustworthiness before publishing."

Google's helpful content self-assessment lists warning signs that fit a lot of AI blog output: "Are you using extensive automation to produce content on many topics?" and "Are you mainly summarizing what others have to say without adding much value?" Speed is not substance. A firm that publishes twenty generic posts a month has produced twenty pages of summary, and Google's own questions point at that pattern. A smaller number of posts, each built from something only the firm knows, fits that guidance better.

Where AI helps and where it does not

The useful division is between tasks that organize knowledge the firm already has and tasks that would require the AI to supply knowledge it does not have.

The pattern in the table is consistent. AI is most useful when the firm has already supplied the substance, and least useful when it is asked to supply the substance itself, which is also where the risk sits.

Keep client information out of the prompt

A blog post rarely needs client information, but drafting habits can pull it in. Opinion 512 says that for self-learning tools, "a client's informed consent is required prior to inputting information" relating to the representation, and that "merely adding general, boiler-plate provisions to engagement letters purporting to authorize the lawyer to use GAI is not sufficient."

Make the safe path the easy one. Before anything goes into a tool, the person drafting should ask, "Is anything in this prompt from a client file?" Write the firm's AI policy so that marketing prompts use fictional facts and approved firm information only, and name the tools staff and vendors may use.

Vendors deserve the same scrutiny. If an agency drafts the firm's posts with AI, ask which tools it uses, what it puts into them, and how it keeps one client's material separate from another's. The supervision language in Opinion 512 is a reasonable basis for those questions.

A fictional example: two ways to draft the same post

Consider a fictional elder law firm in Michigan that wants a post on what to bring to a first meeting about long-term care planning. This is an invented example.

Approach one: a staffer asks a chatbot for a blog post on the topic. The draft is fluent, lists documents any firm might request, and includes a confident statement about eligibility rules that nobody at the firm has checked. The reviewing attorney spends an hour correcting it and still finds it sounds like nobody in particular.

Approach two: the attorney spends twenty minutes answering five questions on a recorded call, covering what she asks for, why, and what clients most often forget. The AI organizes that transcript into a draft. The attorney's review takes fifteen minutes, because the substance was hers to begin with. The firm's knowledge came first and the tool only shaped it, which is the order that keeps anyone, human or software, from having to invent what the firm knows. The transcript also outlasts the post: it can feed an FAQ answer, an intake script, or a future post on a related question.

Give the draft the firm's knowledge first

The second approach is a process, not a trick. It needs a place where approved firm knowledge lives, a record of how the attorneys explain things, and a list of the questions clients actually ask. The page on making firm content sound like the attorney covers the voice side, and how advertising rules apply to blog posts and AI content covers the review side. The article on AI content that still sounds like your business includes a worked example.

Start small: one attorney, one practice area, five questions, one recorded call. If the drafts that come out of it are faster to review and sound more like the firm, the process has earned a second round.

Hudson South's four steps follow that order. Authority Asset gathers the firm's approved knowledge and evidence. Hudson Twenty-One records how the firm communicates across 21 dimensions so the writing sounds like it. Market Intelligence researches the client questions and the market. AEO Creation turns those inputs into answer-first content, which the firm's attorneys then review. Our law firm content guide lists the related questions.

Put your firm's knowledge ahead of the tool

On a free 30-minute call, Hudson South walks through how your blog compares with the firms prospective clients are reading.

Sources

1. Formal Opinion 512: Generative Artificial Intelligence Tools. American Bar Association Standing Committee on Ethics and Professional Responsibility. Read 2026-10-15.
2. Google Search's guidance on using generative AI content on your website. Google Search Central. Read 2026-10-15.
3. Creating helpful, reliable, people-first content. Google Search Central. Read 2026-10-15.