United States (national bar association) · ethics opinion
ABA Formal Op. 512: what it asks, and the records to keep
Opinion 512 is the ABA's first formal ethics opinion on generative AI, applying existing Model Rules on competence, confidentiality, communication, candor, supervision and fees to lawyers' AI use. It is advisory, but state bars and courts widely cite it as the baseline for how lawyers may use AI.
01 / What it asks
What ABA Formal Op. 512 asks
Lawyers who use generative AI tools in practice, and the partners and supervisors responsible for their firms' conduct.
- 01Have a reasonable understanding of a specific AI tool's capabilities and limits before using it (Rule 1.1).
- 02Independently review AI output to a degree suited to the tool and task before relying on it.
- 03Get a client's informed consent before entering their confidential information into a self-learning AI tool; boilerplate engagement-letter consent is not enough (Rule 1.6).
- 04Tell clients about AI use when they ask, when it affects fees, or when it influences significant decisions (Rule 1.4).
- 05Check AI output for false statements and fabricated citations before submitting to a court (Rules 3.1, 3.3, 8.4(c)).
- 06Set firm AI policies, train staff and vet vendors (Rules 5.1, 5.3); bill only reasonable fees for actual time (Rule 1.5).
02 / AI work
Where it touches AI work
The opinion deals only with generative AI tools, such as drafting, research and summarization assistants. It treats AI like other technology and nonlawyer assistance: allowed, but with the lawyer fully responsible for the result.
03 / Dates
The dates that matter
ABA House of Delegates adopts Resolution 604 on AI accountability
Mata v. Avianca sanctions for fabricated AI citations
Formal Opinion 512 issued
04 / Enforcement
Who enforces it
None directly; ABA opinions are advisory. State disciplinary authorities enforce the underlying state rules, and courts may sanction filings.
05 / Records
Records an organisation keeps
What an auditor or an inquiry usually asks to see. Your obligations depend on your role and your facts.
- 01Documented evaluation of each AI tool's terms and data handling
- 02Client informed-consent records for confidential inputs
- 03Verification notes for AI-assisted work product
- 04Firm AI policy and training attendance records
Optional evidence · where Planisphere fits
Keep a record anyone can check
When your team meets a duty, Planisphere can seal the event it reports as a receipt.
A disclosure shown, a review signed, a model change approved: the event your system reports.
The event becomes a sealed receipt, dated and attributed to whoever is signed in.
Paste the receipt into Verify. No account or key.
A record supports your compliance evidence process. It is not a legal finding or a certification. Systems that call Planisphere's gate API can also hold or refuse an action before it runs.
QUESTIONS
The questions that come next
What does ABA Formal Opinion 512 require of lawyers using AI?+
Opinion 512, issued in 2024, does not ban generative AI; it applies six existing Model Rules duties to it: competence (Rule 1.1), including understanding a tool's capabilities and limits and checking its output; confidentiality of client information (Rule 1.6); communication with clients (Rule 1.4); candor toward the tribunal (Rule 3.3); supervision of lawyers and staff (Rules 5.1 and 5.3); and reasonable fees (Rule 1.5). The opinion is advisory. The binding rules are each state's version of the Model Rules, enforced by state disciplinary authorities and, for court filings, by judges. Read the full answer
Do I need a client's consent to put their info into an AI tool?+
Sometimes. Under Opinion 512 and Rule 1.6, a lawyer must first assess the risk that client information entered into a generative AI tool will be disclosed or accessed by others. For self-learning tools, which may use inputs to train the model or expose them to other users, the opinion calls for the client's informed consent before input, and a generic clause in an engagement letter does not suffice. Tools that keep inputs confidential and do not train on them carry less risk, but the lawyer must still check their terms, data retention and security. Read the full answer
How do I show competence (MR 1.1) in the AI tools my firm uses?+
By showing you understood each tool and checked its work. Opinion 512 reads Rule 1.1 to require a reasonable understanding of a specific tool's capabilities and limitations, including its tendency to produce false or invented content, and independent review of its output suited to the tool and task. Useful records include the tool's terms and your testing notes, firm AI policies and training, and evidence that citations and facts in AI-assisted work were verified before use. Opinion 512 also says time spent learning a tool generally cannot be billed unless the client asks for it under a clear agreement. Read the full answer
SOURCES
Go back to the source
American Bar Association Standing Committee on Ethics and Professional Responsibility
American Bar Association
American Bar Association
CHANGE LOG
What changed, and when
Rebuilt from primary sources; facts checked on this date.
Related: ABA Formal Op. 512 for law AI workflows · ABA Model Rules