AI Unauthorized Practice

State of play.

• ABA Formal Opinion 512 and its state-bar adoptions are drawing criticism for internal inconsistency. Mississippi Ethics Opinion No. 267, adopted verbatim from ABA Formal Opinion 512, permits reduced verification for familiar legal-specific AI tools — a standard critics argue is contradicted by the Stanford hallucination data the opinion itself cites (17-33% hallucination rates across leading legal AI platforms) .
• Client-side AI misuse is generating a distinct privilege-waiver risk. In United States v. Heppner (S.D.N.Y.), a federal court held that AI-generated documents created using Claude were not privileged because the tool was not a lawyer and was not used at counsel's direction — a ruling that operationalizes the waiver risk for client-facing practice .
• Sanctions for AI-generated fake citations remain active, while judicial AI use is accelerating. Courts continue to sanction attorneys for hallucinated citations; simultaneously, a reported 60% of judges are using AI tools themselves — a structural asymmetry that has no settled ethical framework .
• State legislatures and the FTC are tightening the regulatory perimeter around AI impersonation of licensed professionals. Pennsylvania, New York, and other states have enacted restrictions on AI impersonation of lawyers; the FTC has pursued injunctions against "robot lawyers" .
• For counsel advising law firms or managing client relationships, the practical baseline is that ethics opinions set a permissive floor — not a safe harbor — and firms without independent governance infrastructure (verification protocols, audit trails, explicit client AI-use instructions) carry both malpractice and privilege-waiver exposure that bar guidance alone does not resolve.

Where things stand.

• ABA Formal Opinion 512 is the governing ethics baseline, but it is contested. The opinion establishes duties of competence, confidentiality, output verification, reasonable billing, and informed consent for AI use. Its permission to reduce verification for familiar tools is the live fault line — a Stanford study cited within the opinion found 17-33% hallucination rates in leading legal AI systems, which critics argue makes experience-based verification shortcuts indefensible .
• State bar adoptions are fragmenting the verification standard. Mississippi adopted ABA Formal Opinion 512 verbatim as Ethics Opinion No. 267. Whether other states adopt, modify, or reject the reduced-verification permission will determine whether a national standard emerges or the landscape splinters .
• Consumer AI platforms create privilege-waiver exposure for both attorneys and clients. Uploading privileged documents — draft agreements, memos, work product — into ChatGPT, Claude, or similar platforms exposes confidential information to third parties with no confidentiality obligations. Heppner provides the judicial anchor for that waiver theory, and the client management dimension is now explicit in practitioner advisories .
• ABA Model Rule 1.6(c) is the operative confidentiality hook. Privacy toggles and similar safeguards in consumer AI platforms do not satisfy the ethical standard for preventing unintended disclosure; inputting client data into public AI systems without enterprise-grade data agreements is a Rule 1.6 exposure .
• Sanctions for AI-generated hallucinations are documented at the highest levels. Chief Justice Roberts cited AI-generated fake citation incidents in his 2023 Annual Report; sanctions continue in federal courts .
• AI impersonation of licensed professionals is a regulated category. The FTC has pursued injunctions against "robot lawyers"; Pennsylvania and New York have enacted statutes restricting AI impersonation of licensed professionals. The regulatory perimeter is tightening around unauthorized practice by AI systems, not just by humans using AI .
• Judicial AI use is outpacing the ethical framework governing it. Approximately 60% of judges report using AI tools, while attorneys continue to face sanctions for AI-assisted work product errors — a structural asymmetry with no current resolution .

Latest developments.

• National Law Review and Varnum LLP advisory warns clients against uploading privileged documents to consumer AI platforms, citing United States v. Heppner (S.D.N.Y.) as judicial backing for privilege waiver; the pieces also flag that AI models tend to validate user assumptions rather than provide objective legal analysis, adding a reliability dimension to the client-management problem; FTC and state-law restrictions on AI impersonation of lawyers noted as tightening regulatory perimeter .

Active questions and open splits.

• Does experience with a legal AI tool justify reduced verification? ABA Formal Opinion 512 and Mississippi Ethics Opinion No. 267 say yes — but the Stanford hallucination data they cite says the premise is unsound. No court has yet tested whether reliance on this guidance defeats a malpractice or sanctions claim .
• What is the scope of privilege waiver when a client uploads attorney work product to a consumer AI platform? Heppner establishes the waiver theory for AI-generated documents not created at counsel's direction, but the outer boundary — how much client-side AI use taints the privilege — is unsettled .
• Does Rule 1.6(c) require enterprise data agreements before any AI use involving client information? Practitioner advisories treat consumer AI platforms as categorically inadequate; bar opinions have not drawn that bright line. The gap between practitioner warnings and formal ethics guidance is live .
• What ethical framework governs judicial AI use? With approximately 60% of judges using AI tools and no parallel sanctions regime, the asymmetry between judicial and attorney AI use is unaddressed. Whether judicial AI use in decision-making triggers disclosure obligations is an open question .
• Where does AI legal assistance end and unauthorized practice begin? The FTC's "robot lawyer" injunctions and state impersonation statutes define one edge; the boundary for AI tools that provide legal information short of representation remains contested across jurisdictions .
• Will the ABA revise Formal Opinion 512 in response to criticism? The opinion's reduced-verification permission is under documented pressure. Whether the ABA issues supplemental guidance — or whether courts and disciplinary bodies simply override it through sanctions — is the near-term resolution path .

What to watch.

• Whether the ABA responds to the Formal Opinion 512 criticism with supplemental guidance on verification standards, particularly for rapidly evolving AI systems.
• Additional state bar ethics opinions adopting, modifying, or rejecting the ABA's reduced-verification permission — Mississippi is the first but unlikely to be the last.
• Sanctions decisions and malpractice filings that test whether reliance on ABA or state bar AI guidance constitutes a defense.
• Further judicial decisions applying or extending Heppner's privilege-waiver theory to client-side AI use in civil matters — the S.D.N.Y. ruling is the current anchor but its reach into civil privilege disputes is untested.
• FTC and state AG enforcement actions against AI platforms marketing legal services, which will sharpen the unauthorized-practice perimeter.
• Whether any court or disciplinary body addresses the judicial AI use asymmetry through disclosure requirements or recusal standards.