Are ABA Formal Opinions Binding? What Opinion 512 Means for Your AI Workflow

No, ABA formal opinions are not binding. Not on a court, not on a disciplinary board, not anywhere, until a state authority chooses to give them weight. ABA formal opinion authority is persuasive only: the document is the ABA's reading of its Model Rules, and a court or state bar may follow it, cite it, or ignore it. ABA Opinion 512 is not binding by itself either.

A solo practitioner posted a version of this question to a legal forum last year, and it is the right instinct dressed up in the wrong vocabulary: "ABA Formal Opinion 512 isn't binding, so why does everyone keep telling me I have to follow it?"

The short answer is that they are both right and you should still worry. Opinion 512, the ABA's July 2024 guidance on generative AI for lawyers, binds nobody on its own. If you want the rule-by-rule duty detail, we wrote that companion piece on what Opinion 512 actually requires.

This post is about the prior question: what kind of authority is this document, and why does "not binding" turn out to be cold comfort.

The five Model Rule duties ABA Opinion 512 maps onto AI use.

TL;DR

• ABA formal opinions are persuasive authority, not binding law. The ABA is a private membership organization. Its Model Rules are not law in any state until that state's supreme court adopts them, and its ethics opinions interpret those Model Rules without the force of a statute or a court ruling.
• Opinion 512 creates no new duties. It applies six existing Model Rules to AI. Even if you ignore the document, the underlying duties (competence, confidentiality, candor, supervision, fees, communication) already bind you through your own state's adopted rules.
• States do not wait for the ABA. Florida issued its own AI ethics opinion six months before 512. Pennsylvania and Philadelphia went further than 512 on disclosure.
• By mid-2026 the real authority is a patchwork, roughly a dozen state and local bars with their own generative-AI guidance, most tracking 512, some exceeding it.
• The binding-versus-persuasive debate is a distraction. The live compliance variable is your tool and your workflow, not which document gets labeled "law."

Part of our legal AI verification and hallucination guide series.

For related verification / hallucination / vendor-trust coverage, see ABA Formal Opinion 512 Explained: What Every Lawyer Using GenAI Must Do and "We Do Not Train on Your Data": How to Verify the Claim.

What "Binding" Actually Means Here

Lawyers throw the word "binding" around loosely, so let us be precise, because the precision is the whole point.

The American Bar Association is not a regulator. It is a voluntary professional association. It cannot suspend your license, sanction you, or compel a single thing.

What it produces are Model Rules of Professional Conduct, model because they are a template, an offer to the states, not a code that takes effect on its own. A Model Rule becomes enforceable law only when a state's highest court adopts it (usually the state supreme court, exercising its authority over the bar).

California is the structural outlier here: it never adopted the Model Rules in the ABA's format and built its own rules instead, which is exactly why "the ABA says" lands differently in Sacramento than it does in most other capitals.

Formal opinions sit one layer further out. A formal opinion does not even propose a new rule. It is the ABA's Standing Committee on Ethics and Professional Responsibility saying, "Here is how we read the existing Model Rules as applied to a new situation."

That reading is persuasive: a court or a state bar may find it convincing and cite it, the way a judge cites a respected treatise or a sister-state decision. But nobody is obligated to agree. A disciplinary board in your state can read 512, decide it disagrees with a particular passage, and go its own way.

So when someone tells you Opinion 512 is not binding, they are correct in the narrowest, most technically defensible sense. The mistake is what they do with that fact.

Why "Not Binding" Is the Wrong Place to Stop

Here is the trap. "Not binding" gets mentally filed next to "ignorable," and those are not the same word.

Two things are true at once. First, 512 itself commands nothing. Second, everything 512 describes is already binding on you through a different door.

The opinion does not invent a verification duty or a confidentiality duty. It points at Model Rule 1.1 (competence), Rule 1.6 (confidentiality), Rule 3.3 (candor toward the tribunal), Rules 5.1 and 5.3 (supervision), Rule 1.5 (fees), and Rule 1.4 (communication), and says these apply when you hand work to a machine.

Those rules, in your state's adopted version, already bind you. They bound you before 512 existed. They will bind you if 512 is rescinded tomorrow.

The lawyer in Mata v. Avianca was not sanctioned for violating an ABA opinion. He was sanctioned in June 2023, more than a year before 512, under Federal Rule of Civil Procedure 11, for filing fabricated case citations a chatbot produced.

The duty he breached, candor and a reasonable inquiry into his filings, predates every AI ethics opinion ever written. We traced that whole arc and the verification protocol that prevents it in our piece on AI hallucinations and sanctions. 512 did not create his problem. It described a problem the rules already covered.

This is why "is it binding" is the wrong question. The duties are binding. 512 is just the clearest available map of how they apply to AI, which means a disciplinary board, when it eventually evaluates an AI-related complaint, will reach for 512 as persuasive authority for what reasonable conduct looked like.

Persuasive authority is not nothing. It is the thing the decision-maker cites when they rule against you.

States Do Not Wait for the ABA

If you still think the ABA leads and the states follow, the timeline says otherwise.

Florida issued the first formal ethics opinion. The Florida Bar approved its own ethics opinion on generative AI (Opinion 24-1) on January 19, 2024, roughly six months before the ABA published 512 in July 2024. California had already published "Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law" in November 2023, so the states were out ahead of the ABA on both counts.

Neither waited for national guidance. They read their own rules, saw lawyers already using these tools, and issued direction. That ordering matters: it shows states treat AI ethics as their own jurisdiction to regulate, not a federal-style hierarchy where the ABA legislates and everyone falls in line.

Pennsylvania and Philadelphia went further than 512. The joint formal opinion from the Pennsylvania Bar Association and the Philadelphia Bar Association (2024-200) emphasized competence and, on the question of disclosure, pushed a stricter posture than the ABA: coverage of the opinion described it as leaning toward explicit disclosure of AI use in court submissions. Compare that to 512, which declined to impose a blanket duty to tell clients or courts every time you use AI. Same technology, same year, and a meaningfully stricter rule depending on which side of a state line you practice on.

By early 2025, a running tally kept by LawSites (LawNext) counted roughly a dozen state and local bars with their own generative-AI ethics guidance: California, Florida, New York State, New York City, New Jersey, Pennsylvania, Philadelphia, Kentucky, North Carolina, Texas, Virginia, and the District of Columbia.

The number has only grown since. Most of these track 512's core logic, which is unsurprising given they read the same underlying Model Rules. Some go further on disclosure or on supervision specifics.

The practical takeaway: the authority you actually answer to is not one ABA document. It is your state's version, which has very likely already echoed 512 and may have raised the floor above it. Treat 512 as the national baseline and then go read what your own jurisdiction did, because that is the document a disciplinary board in your state will apply directly.

Try Vaquill AI on your matterBook a demo

What Most People Get Wrong

Two errors, equal and opposite.

The first is the one we have been picking apart: treating "not binding" as "ignorable." That reasoning works right up until a complaint lands, at which point the underlying rule (which is binding) gets enforced and 512 gets cited as the standard of reasonable care. You do not get to argue that an unbounded verification duty did not apply to you because the ABA's interpretation of it was "merely persuasive."

The second error is the inverse: treating 512 as the rule, a self-contained AI statute you can comply with by reading one PDF. It is not. It creates no new duties, it draws almost no bright lines (the required level of verification is "fact-specific" by design), and it does not control in your state.

Lawyers who memorize 512 and stop there can still be blindsided by a state opinion that requires court disclosure 512 never demanded.

The honest framing sits between them. 512 is a floor, articulated nationally, that your own state has probably already built on.

Whatever the label on the document, you still own three things personally: verification of what the tool produces, confidentiality of what you feed it, and honest billing for the time it saves or costs. None of those obligations cares whether you call its source "binding."

So What Does This Mean for Your AI Workflow

Strip away the authority taxonomy and the operational reality is refreshingly concrete. The compliance variable you control is not which document is law. It is the tool you pick and the workflow you wrap around it.

Verification. Because every version of the competence duty leaves "how much checking is enough" deliberately open, the cheapest way to stay defensible is to make verification fast.

A tool that retrieves from real opinions and shows you the source it relied on (retrieval-augmented generation, not a model reciting from training data) lets you open the citation, read the passage, and confirm it says what the tool claims, in seconds. A tool that conjures a citation from nothing forces you to redo the entire research task by hand.

The first is how grounded research surfaces are supposed to work; the second is how Mata happens. We laid out the step-by-step routine in how to verify AI legal citations before filing.

Confidentiality. Both 512 and most state opinions converge on the same instinct: be careful what client information you feed into a self-learning system. That makes your vendor's data posture a compliance question, not just a procurement one.

Whether a tool trains on your inputs, how long it retains them, and what its data-processing terms actually say all matter. Any serious vendor should let you read its security and DPA terms before you trust it with privileged facts.

Primary sources. A surprising amount of "AI compliance" is just the old duty to check primary authority, dressed in new clothes. When you are verifying that a statute says what a brief claims, you want the actual text of the U.S. Code, the CFR, or a state code, not a paraphrase.

A public statutes API that exposes USC, CFR, federal rules, and 50-state codes for programmatic search and retrieval of primary text fills that gap. It will not write your brief, but it answers "what does the statute actually say" without a paraphrase in between.

(Florida and New York lawyers, in particular, will recognize this as the same instinct behind their own state bars' AI guidance: the burden of confirming primary authority does not move just because a tool drafted the paragraph.)

Fees. The quietest duty in the whole 512 cluster is Rule 1.5. You generally cannot bill a client for the time you spend learning a new tool, and efficiency gains from AI raise honest questions about flat fees versus hourly billing.

That is a workflow and engagement-letter problem, not a software feature, and it is the one most likely to surface in a fee dispute rather than a malpractice claim. Worth a line in your engagement letter before it is worth a paragraph in a grievance.

The Real Question to Ask Instead

Stop asking whether ABA Formal Opinion 512 is binding. It is not, and the answer leads you nowhere useful.

Ask instead: what has my state's highest court adopted, what has my state bar said about AI specifically, and have I picked tools and built a workflow that make verification cheap, confidentiality defensible, and billing honest?

Those are the questions a disciplinary board will effectively be asking on your behalf, after the fact, with 512 open on the desk as persuasive authority. Better to answer them first.

The label on the document is a distraction. The duties under it are not.

FAQ

Are ABA formal opinions binding?

No. The American Bar Association is a private membership organization, not a regulator, so its formal opinions are persuasive authority, not law. An opinion becomes something you actually answer to only when your state's highest court or bar gives the underlying rule weight. A court or disciplinary board can read an ABA opinion, disagree with it, and rule the other way.

What is ABA formal opinion authority?

It is persuasive, the same kind of weight a court gives a respected treatise or a sister-state decision. A formal opinion is the ABA Standing Committee on Ethics and Professional Responsibility explaining how it reads the Model Rules in a new situation. Decision-makers cite it when they find it convincing, but nobody is obligated to follow it.

Is ABA Opinion 512 binding?

No. Opinion 512, issued July 29, 2024, creates no new duties and binds no one by itself. What binds you are the duties it points to (competence, confidentiality, candor, supervision, communication, fees) as adopted in your own state's rules of professional conduct. Those rules bound you before 512 and would survive its repeal.

What actually binds a lawyer instead of an ABA opinion?

Your state's rules of professional conduct, as adopted and enforced by your state's highest court, plus any ethics opinion or court ruling your own jurisdiction has issued. The ABA Model Rules are a template; they become enforceable only after a state adopts them. California, for example, never adopted the Model Rules in the ABA's format and wrote its own.

Do all states follow ABA Opinion 512?

No. States may follow it, diverge from it, or go further. By mid-2026 roughly a dozen state and local bars had published their own generative-AI guidance, most tracking 512's core logic because they read the same Model Rules, some stricter on disclosure. Pennsylvania and Philadelphia's joint opinion pushed a stricter disclosure posture than 512.

Does Opinion 512 require lawyers to disclose AI use to clients or courts?

Not in every case. 512 treats disclosure as fact-specific rather than a blanket rule, and it declined to impose an across-the-board duty to tell clients or courts every time you use AI. Some states have gone further, so the safe move is to check your own jurisdiction's guidance, which is the document that applies to you directly.

A closing takeaway

The label on the document is the wrong fight. The duties under it (verification, confidentiality, supervision, fees) are the work. Pick tools whose data flow you can defend in writing, and a workflow that makes cite-checking cheap enough that you actually do it.

Making verification cheap is the practical part, and it is why Vaquill AI grounds its legal research in real opinions with citation verification, so you can open the source and confirm it says what the tool claims in seconds.

Want a research surface you can defend in writing? Start a free Vaquill AI trial, or see how citation-verified research works.

Arshita Anand

Co-Founder & CEO · Attorney

Arshita leads product and strategy at Vaquill, building the legal AI suite that solo, small-firm, and in-house US lawyers use to run a matter end to end.