ABA Model Rule 1.1 and AI: What Competence Requires of Attorneys Using AI Tools

The Rule and the Technology Comment

ABA Model Rule 1.1 states that a lawyer shall provide competent representation, which requires "the legal knowledge, skill, thoroughness and preparation reasonably necessary for the representation." That language has not changed. What changed in 2012 was Comment 8, which added that competence includes keeping abreast of "changes in the law and its practice, including the benefits and risks associated with relevant technology."

Comment 8 predates generative AI by a decade, but its language is deliberately forward-looking. The ABA Standing Committee on Ethics and Professional Responsibility has consistently interpreted it to cover AI tools as they have entered legal workflows — and several state bars have issued their own formal opinions building directly on this foundation.

What "Technological Competence" Means for AI

The competence obligation under Rule 1.1 does not require attorneys to become software engineers. It requires them to understand, at a functional level, how the tools they deploy behave — including their documented failure modes. For AI tools used in legal research, drafting, or contract review, that translates into several concrete obligations.

• Understanding that large language models can fabricate citations, case names, and statutory text — and building review steps into any workflow that relies on AI-generated legal authority.
• Knowing the difference between AI tools that retrieve from a verified legal corpus (RAG-based systems like Westlaw CoCounsel or Lexis+ AI) and those that generate text from training data without live retrieval, where hallucination risk is materially higher.
• Being able to evaluate whether an AI tool's output is reliable enough for the specific task — a standard that varies by task type, stakes, and the tool's documented accuracy profile.
• Not delegating final judgment on legal conclusions, strategy, or client advice to an AI system, regardless of how confident the output appears.

The ABA's Formal Opinion 512 (issued in July 2024) addressed generative AI directly and confirmed that Rule 1.1 requires lawyers to understand the technology they use sufficiently to assess its output. The opinion noted that this does not mean attorneys must personally master every underlying model, but it does mean they cannot treat AI output as equivalent to verified legal authority without independent review.

How State Bars Have Interpreted the Competence Obligation

By mid-2026, more than 25 state bars have issued AI-specific ethics opinions or formal guidance addressing attorney obligations. Most track the ABA framework, but several have gone further in specifying what competence requires in practice.

The Supervision Dimension: Rule 5.1 and 5.3

Rule 1.1 competence does not operate in isolation. The ABA and most state bars have read AI obligations as engaging Rules 5.1 and 5.3 simultaneously — the rules governing supervision of other lawyers and nonlawyer assistants respectively.

ABA Formal Opinion 512 treats AI tools as analogous to nonlawyer assistants for supervision purposes. A supervising attorney who deploys an AI tool in a workflow bears responsibility for the output that workflow produces, in the same way a partner bears responsibility for work delegated to a paralegal. The supervision obligation requires establishing adequate review procedures — not just using the tool and hoping the output is correct.

In practice, this means law firms deploying AI tools need defined review protocols at the workflow level, not just general AI policies. An associate who uses an AI legal research tool to identify case authority still needs to verify each citation independently before it appears in a brief. A partner who reviews that brief without checking the AI-generated citations has potentially failed their own supervision obligation.

Where Competence Failures Have Led to Sanctions

The practical stakes of Rule 1.1 in the AI context became concrete in 2023 when the Southern District of New York sanctioned attorneys in Mata v. Avianca for filing a brief containing AI-generated citations to cases that did not exist. The court found that the attorneys had failed to verify the citations before filing — a straightforward competence failure that the court treated as sanctionable conduct.

That case has since been cited in multiple bar ethics opinions as a concrete illustration of what Rule 1.1 requires. The lesson is not that AI cannot be used for legal research — it is that AI-generated legal authority must be independently verified before it is relied upon in any filing or client advice.

Confidentiality and Competence: The Data Input Problem

Rule 1.1 competence intersects with Rule 1.6 confidentiality in AI workflows because using an AI tool requires inputting client information. An attorney who does not understand how a specific tool handles submitted data — whether it retains prompts, uses them for model training, or transmits them to third parties — cannot make a competent judgment about whether using that tool is consistent with their confidentiality obligations.

ABA Formal Opinion 512 addressed this directly, noting that attorneys must review the terms of service and data handling policies of any AI tool before inputting client confidential information. Tools marketed to the legal sector increasingly offer zero-retention or isolated-instance deployments specifically to address this requirement, but attorneys bear the obligation to verify the actual data handling model rather than assume it.

Disclosure Obligations: When Must Clients Be Told?

There is no universal rule requiring disclosure of AI use to clients, but the competence and confidentiality framework creates conditional disclosure obligations in specific circumstances.

• If using an AI tool requires transmitting client confidential information to a third-party service, the attorney may need client consent under Rule 1.6(a) before proceeding — particularly where the engagement agreement does not contemplate AI-assisted work.
• If a fee arrangement is affected by AI use (e.g., a task that previously took four hours now takes thirty minutes with AI assistance), some state bars have indicated that billing transparency obligations may require disclosure.
• California's 2023 practical guidance explicitly recommends disclosure where AI use is material to the representation, though it stops short of making this a mandatory requirement in all cases.
• Several state bars, including New York's task force, have noted that retainer agreements should address AI use prospectively rather than requiring case-by-case disclosure decisions.

Practical Compliance: What the Rule Requires in Operation

Translating Rule 1.1 obligations into day-to-day AI use requires more than a general awareness that AI can make mistakes. The bar guidance issued through 2026 points to a set of concrete operational requirements.

What Rule 1.1 Does Not Require

The competence obligation is sometimes read more broadly than the rule supports. Several things Rule 1.1 does not require are worth stating explicitly.

• Attorneys are not required to use AI tools. Rule 1.1 requires competence in the tools actually used, not adoption of any particular technology.
• Attorneys are not required to become experts in how large language models work at a technical level. The standard is functional understanding sufficient to assess output quality — not the ability to explain transformer architecture.
• There is no affirmative duty to adopt AI tools to increase efficiency, even where AI might make a task faster or cheaper. The competence obligation runs to quality of representation, not speed.
• Rule 1.1 does not itself require disclosure of AI use to clients in every case. Disclosure obligations arise from the intersection with Rules 1.4 and 1.6 in specific factual circumstances.

ABA Formal Opinion 512: Key Holdings

ABA Formal Opinion 512, issued July 29, 2024, is the ABA's most direct treatment of generative AI under the Model Rules. Its holdings under Rule 1.1 can be summarized as follows.

Lawyers who use generative AI must understand the technology's benefits and risks, including the risk that the AI may generate inaccurate information. Lawyers must review the output of generative AI tools and may not delegate their professional judgment to such tools.

ABA Formal Opinion 512 (July 29, 2024)

The opinion also addressed the supervision obligation, confirming that Rule 5.3 applies to AI tools in the same way it applies to nonlawyer staff — meaning that the responsible attorney is accountable for the work product the AI generates, and must have adequate review procedures in place to catch errors before they reach clients or courts.

Primary Source and Verification

The authoritative text of ABA Model Rule 1.1 and its comments is maintained by the ABA Center for Professional Responsibility. ABA Formal Opinion 512 is available through the ABA's ethics opinions database. State bar opinions referenced in this entry should be verified directly with each issuing bar, as guidance is frequently supplemented or superseded.

This entry was last confirmed against available primary sources in June 2026. Attorneys tracking compliance obligations should monitor their state bar's ethics opinion database for updates, as the volume of AI-specific guidance has been increasing and several state bars have indicated that additional opinions are in progress.