TL;DR: Yes - using AI in legal practice is ethical when you do it competently and keep a human in control. ABA Model Rule 1.1 and its Comment 8 require technology competence, and ABA Formal Opinion 512 (July 2024) maps generative AI onto existing duties: competence, confidentiality, client communication, candor, supervision, and reasonable fees. AI is a tool; the lawyer remains responsible for the work.
The question is asked nervously in partner meetings and CLE sessions alike: is it even ethical to use generative AI for legal work? The honest answer is that ethics rules do not ban the technology - they govern how you use it. The American Bar Association has been clear that competent, supervised use is permissible, and several state bars have followed with guidance of their own. This article explains the rule at the center of the debate, ABA Model Rule 1.1, walks through Formal Opinion 512, and translates the duties into practical habits.
Is it ethical to use AI in legal practice?
Start with the conclusion, because it is not in serious doubt: using AI is ethical, provided you use it competently and supervise its output. No major bar authority has prohibited generative AI for legal work. What they have done is remind lawyers that existing duties - to be competent, to protect confidences, to be candid with courts, and to charge reasonable fees - apply with full force to AI-assisted work. The technology does not create a loophole, and it does not create a new category of misconduct; it is measured against the standards that already govern everything a lawyer does.
The nuance is that ethical use is not automatic. A lawyer who pastes privileged client data into a public chatbot, files its output unchecked, or bills a client for hours the tool actually saved may well breach the rules - not because AI is involved, but because of how it was used. The same tool, handled with care, is perfectly proper. Ethics, in other words, is about the practitioner's judgment, not the presence of the machine. The rest of this guide is about what that judgment requires.
What does ABA Model Rule 1.1 require?
ABA Model Rule 1.1 states the foundational duty of competence: a lawyer shall provide competent representation, which requires the legal knowledge, skill, thoroughness, and preparation reasonably necessary for the representation. It is the first rule in the Model Rules for a reason - everything else builds on the premise that a lawyer is competent to handle the matter. The full text and official comments are published by the American Bar Association.
Competence has always been understood to evolve with practice. A lawyer cannot rely on the knowledge that was current at admission; the duty is ongoing. As legal work has come to depend on technology, the competence standard has stretched to cover the tools as well as the law - and that is where the now-famous Comment 8 enters the picture.
Comment 8 and the duty of technology competence
Comment 8 to Rule 1.1 provides that, to maintain the requisite knowledge and skill, a lawyer should keep abreast of changes in the law and its practice, including the benefits and risks associated with relevant technology. Adopted by the ABA in 2012 and since incorporated by a large majority of states, it established what is widely called the duty of technology competence. For AI, the implication is direct: a lawyer who uses a generative tool is expected to understand, at least in general terms, both what it can do and how it can fail - including its tendency to hallucinate.
This cuts against two opposite errors. The lawyer who refuses to learn anything about AI may fall short of competence as the technology becomes standard; the lawyer who adopts it uncritically, trusting fluent output without grasping the risks, falls short in the other direction. Comment 8 asks for informed use - enough understanding to deploy the tool where it helps and to guard against the ways it misleads. Our explainer on what legal AI is is a reasonable place to build that baseline understanding.
What does ABA Formal Opinion 512 say about generative AI?
In July 2024, the ABA Standing Committee on Ethics and Professional Responsibility issued Formal Opinion 512, its first comprehensive guidance on generative AI. Rather than inventing new rules, the opinion maps the technology onto duties already in the Model Rules and explains how each applies. It is measured in tone: AI can make legal services more efficient and accessible, the committee acknowledges, while raising real questions a careful lawyer must address before and during use.
The opinion touches a cluster of duties. Competence (Rule 1.1) requires a reasonable understanding of the tool. Confidentiality (Rule 1.6) requires caution before inputting client information into systems that might expose or reuse it. Communication (Rule 1.4) may require telling clients how AI is being used in their matter. Candor toward the tribunal (Rule 3.3) requires verifying that anything submitted to a court is accurate - the lesson of the Mata v. Avianca sanctions. Supervision (Rules 5.1 and 5.3) requires partners and supervising lawyers to ensure that subordinates and non-lawyer tools are used consistently with the rules. And fees (Rule 1.5) must remain reasonable, which means a lawyer generally should not bill for time that AI saved as though it had been spent. State bars have begun issuing their own opinions in a similar spirit, which we track in our roundup of state bar AI ethics opinions.
Which ethics rules are engaged when you use AI?
It helps to see the duties side by side, each paired with the concrete obligation it imposes on AI-assisted work. The table below summarises the core rules that Formal Opinion 512 and related guidance bring to bear. It is a map, not a substitute for reading the rules and any opinions in your own jurisdiction.
| ABA Model Rule | Duty | What it requires when using AI |
|---|---|---|
| Rule 1.1 (and Comment 8) | Competence | Understand the benefits and risks of the AI tool you use, including how it can err |
| Rule 1.6 | Confidentiality | Do not expose client information to tools that may store, train on, or reuse it without protection |
| Rule 1.4 | Communication | Where appropriate, tell the client how AI is used in their matter |
| Rule 3.3 | Candor to the tribunal | Verify that every authority and statement submitted to a court is accurate |
| Rules 5.1 and 5.3 | Supervision | Ensure lawyers and AI tools alike are used consistently with the rules |
| Rule 1.5 | Reasonable fees | Do not bill for time the tool saved; charge for value and work actually performed |
Is AI just a tool, or does it change the lawyer's role?
It is fashionable to say AI is just a tool, and in the sense that matters for ethics, that is right: responsibility does not transfer to the software. A drafting assistant is, in this respect, like a junior associate or an outside vendor - it can do work, but the lawyer who relies on it owns the result. When a tool produces a fabricated citation, the rule violated is the lawyer's, not the model's, because the model is not a member of the bar.
But calling it just a tool understates how much the lawyer's day-to-day role shifts. As AI absorbs more of the mechanical work - first-draft research, document review, summarisation - the lawyer's value concentrates in the parts a tool cannot do: judgment, strategy, client counsel, and the verification that keeps the work honest. That is not a diminished role; if anything, it raises the premium on the distinctly legal skills. The healthiest framing is that AI changes where a lawyer spends time, not whether the lawyer is responsible. We develop that idea further in human-in-the-loop legal AI.
How do you use AI ethically in day-to-day practice?
The rules translate into a short set of habits that are easy to state and worth enforcing. Understand your tools well enough to know how they fail. Protect confidentiality by choosing systems that do not train on your data and that control access, and think carefully before inputting client information anywhere. Verify every citation and factual claim against a primary source before it reaches a filing or a client. Supervise AI output the way you would a junior's work. Be transparent with clients where the rules or good practice call for it. And keep fees tied to value and work actually done, passing on efficiency rather than hiding it.
For firms, these habits are worth writing into a policy so they are applied consistently rather than left to individual discretion. A clear governance framework - approved tools, confidentiality rules, verification standards, and supervision expectations - turns ethical AI use from an aspiration into a routine. We lay out how to build one in our guide to AI governance for law firms, and the verification piece is covered step by step in how to verify AI legal research.
Where does Judicio fit?
No tool can discharge a lawyer's ethical duties, and Judicio does not claim to - outputs are not legal advice, and the human keeps judgment. What a well-designed tool can do is make compliance easier. On competence and candor, Judicio's Legal Research cites every answer to the exact page and quoted passage, with deterministic citation labels and web sources archived as permanent PDFs, so verifying authority before you file is a matter of seconds rather than an afternoon. That directly supports the Rule 3.3 duty that Mata made vivid.
On confidentiality, the design choices map onto Rule 1.6: Judicio does not train on your data, hosts on Google Cloud Platform, and provides role-based access with a full audit trail, so client material stays controlled and traceable. On supervision, the same audit trail makes it possible to see what was run and by whom. None of this replaces a lawyer's judgment - it lowers the friction of exercising it well. You can test that on your own matters with a 7-day free trial of 500 credits, no credit card required, and contact us if you want to discuss your firm's policy.
This article is general information about legal ethics and technology, not legal advice; consult the rules and ethics opinions in your jurisdiction, and remember that Judicio outputs are not legal advice.