AI Legal Malpractice

State of play.

• The "negligence for non-use" question has arrived as a live professional responsibility issue. A University of Houston law professor has publicly framed failure to use AI as a potential competence violation—directly inverting the current judicial consensus that punishes AI misuse, and signaling that the standard of reasonable diligence may soon require demonstrated AI competence, not just avoidance of AI errors (→ Lawyer Faces Potential Negligence Claim for Failing to Use AI in Legal Work).
• Multi-jurisdictional sanctions are now a routine enforcement pattern, not isolated incidents. Connecticut, Illinois, Arizona, and Tennessee federal courts have all imposed sanctions for AI hallucinations in the same regeneration window—with Connecticut additionally codifying mandatory verification rules carrying case-ending consequences (→ Connecticut Supreme Court orders lawyer CLE after ChatGPT citation errors, Illinois appeals court sanctions lawyer over AI-fabricated citations in Starbucks bias case, Arizona Attorney Maren Bam Faces Sanctions for Bogus AI-Generated Quotes in Employment Case, Tennessee Firm Sanctioned $45K for AI-Generated Fake Citations in Malpractice Suit).
• The Tennessee sanctions order in a malpractice-on-malpractice posture is structurally notable. Reaves Law filed a malpractice suit against Baker Donelson—then got sanctioned $45,000 for submitting AI-hallucinated citations in that very proceeding, with the court referring the firm to the Tennessee Board of Professional Responsibility and circulating the order to other district judges (→ Tennessee Firm Sanctioned $45K for AI-Generated Fake Citations in Malpractice Suit).
• Nine-figure malpractice exposure is materializing from regulatory-advice failures on novel business models. MV Realty's $400M–$1.2B suit against Holland & Knight and the Simpson Thacher Patriot National trial both turn on whether counsel adequately anticipated foreseeable regulatory and market risks—a duty-scope question with direct implications for firms advising fintech, real estate, and capital-markets clients .
• For counsel advising law firms or individual practitioners, the practical baseline is that AI governance failures now carry simultaneous exposure in sanctions proceedings, bar discipline referrals, and malpractice suits—and the emerging "negligence for non-use" framing means the window for building defensible AI protocols is narrowing from both directions.

Where things stand.

• Verification of AI-generated citations is a nondelegable duty across jurisdictions. Courts in Tennessee, Arizona, Illinois, Connecticut, and multiple federal circuits have all treated submission of hallucinated citations as a Rule 11 or professional conduct violation regardless of how the work was generated (→ Tennessee Firm Sanctioned $45K for AI-Generated Fake Citations in Malpractice Suit, Arizona Attorney Maren Bam Faces Sanctions for Bogus AI-Generated Quotes in Employment Case, Illinois appeals court sanctions lawyer over AI-fabricated citations in Starbucks bias case, Connecticut Supreme Court orders lawyer CLE after ChatGPT citation errors).
• Connecticut has codified mandatory AI verification with case-ending sanctions. The Connecticut Appellate and Supreme Courts' Practice Book rule, effective June 23, 2026, requires independent verification of all AI-generated citations, legal authorities, and evidence, with nonsuit or default judgment available for violations—the most structurally significant court-rule development in this cluster .
• Repeat-offender patterns draw escalating responses. Arizona attorney Maren Bam was sanctioned twice by the same court for the same conduct—pro hac vice revocation, brief striking, and bar notification in the first proceeding, with fee sanctions under active consideration in the second (→ Arizona Attorney Maren Bam Faces Sanctions for Bogus AI-Generated Quotes in Employment Case).
• Bar discipline referrals are now a standard judicial response. The Tennessee court's referral of Reaves Law to the Tennessee Board of Professional Responsibility—and circulation of the sanctions order to other district judges—signals that courts are treating AI hallucination sanctions as a disciplinary escalation pathway, not a terminal monetary remedy (→ Tennessee Firm Sanctioned $45K for AI-Generated Fake Citations in Malpractice Suit).
• The "negligence for non-use" framing is entering professional discourse. The counterintuitive argument that failure to adopt AI could itself constitute a competence violation has been raised in academic commentary and is now circulating in practitioner media—creating a doctrinal pincer that firms must account for in AI governance design (→ Lawyer Faces Potential Negligence Claim for Failing to Use AI in Legal Work).
• AI-driven discovery review is categorically distinct from AI-assisted review under emerging doctrine. Earlier corpus holdings establish that exclusive AI delegation in discovery review violates FRCP good-faith obligations—a principle that has not yet been extended to brief drafting or due diligence workflows but presents the same substitution risk.
• Duty-scope in capital transactions and regulatory-advice contexts is under active litigation. The Simpson Thacher trial turns on whether the firm's duty ran to the founder personally or only to the corporate client; the Holland & Knight suit turns on whether counsel was obligated to anticipate multi-state AG resistance to a novel fee structure—both will produce jury findings that reshape how firms document client relationships and scope engagement letters .
• AI-driven client solicitation is a distinct TCPA exposure vector. A Texas federal class action targets mass tort firms for deploying AI to generate automated solicitation calls without required consumer consent—the first significant challenge to AI-generated telemarketing in legal intake, proceeding under the Fifth Circuit's Bradford v. Sovereign Pest Control consent standard .

Latest developments.

• Connecticut Supreme Court sanctions Ian Gottlieb of GLG Law LLC for ChatGPT-generated fabricated citations in two landlord-tenant appeals, ordering six additional CLE hours in ethics and law office management—the first Connecticut Supreme Court sanction specifically for AI hallucinations (→ Connecticut Supreme Court orders lawyer CLE after ChatGPT citation errors).
• Illinois appellate court unanimously sanctions an attorney for AI-fabricated citations and misquoted cases in a Starbucks workplace discrimination matter, continuing a pattern of multiple Illinois appellate sanctions for the same conduct (→ Illinois appeals court sanctions lawyer over AI-fabricated citations in Starbucks bias case).
• University of Houston professor Renee Knake Jefferson raises the "negligence for non-use" question in Above the Law's Legal Ethics Roundup, framing AI adoption as a potential component of reasonable diligence rather than a liability risk (→ Lawyer Faces Potential Negligence Claim for Failing to Use AI in Legal Work).
• Texas federal class action filed against mass tort firms for AI-generated automated solicitation calls without TCPA-required consumer consent .
• Tennessee federal court orders Reaves Law Firm to pay Baker Donelson $45,000 for AI-hallucinated citations in a malpractice suit against Baker Donelson itself; court refers firm to Tennessee Board of Professional Responsibility and circulates order to other district judges (→ Tennessee Firm Sanctioned $45K for AI-Generated Fake Citations in Malpractice Suit).
• Arizona federal court weighs additional fee sanctions against Maren Bam for 19 AI-generated citations (5–7 of which existed) and fabricated quotations attributed to Arizona federal judges in a Phoenix Suns discrimination matter—a second sanctions proceeding following August 2025 pro hac vice revocation for the same conduct (→ Arizona Attorney Maren Bam Faces Sanctions for Bogus AI-Generated Quotes in Employment Case).
• Connecticut Appellate and Supreme Courts impose mandatory AI verification rules effective June 23, 2026, with case-ending sanctions—nonsuit or default judgment—available for violations .
• Connecticut attorney Steven A. Schwartz appears before the state Supreme Court to acknowledge unverified ChatGPT use in briefs filed in January 2026, with a sanctions hearing resulting—the same Schwartz previously fined $5,000 in the S.D.N.Y. for the same conduct .
• Simpson Thacher faces a Florida malpractice jury on whether its advice during a 2015 capital raise contributed to Patriot National's "death spiral" bankruptcy; central dispute is whether the firm's duty ran to founder Mariano personally or only to the corporate client .
• MV Realty sues Holland & Knight for $400M–$1.2B alleging defective counsel on the Homeowner Benefit Agreement fee structure that drew multi-state AG enforcement .

Active questions and open splits.

• Does the "negligence for non-use" theory have doctrinal traction? The argument that failing to use AI could itself breach the duty of competence has entered professional discourse but has no judicial backing yet—the question is whether any bar ethics opinion or court will adopt it, and what workflow obligations it would impose (→ Lawyer Faces Potential Negligence Claim for Failing to Use AI in Legal Work).
• What verification protocol affirmatively satisfies Rule 1.1 competence? Courts have defined the violation (submitting unverified AI output) but no court or bar opinion has specified what independent verification requires in operational terms—leaving firms to construct protocols without a safe-harbor standard.
• Does a no-AI office policy insulate the supervising attorney under Model Rule 5.1? Prior corpus holdings establish it does not, but courts have not articulated what enforcement infrastructure—audit trails, workflow controls, attestation requirements—would satisfy supervisory responsibility.
• Will Connecticut's mandatory verification rule become the national template? Connecticut's Practice Book rule is the most structurally significant court-rule development in this cluster; whether other state supreme courts or federal districts adopt comparable mandatory-verification frameworks with case-ending sanctions is the key jurisdictional question .
• Do repeat sanctions in the same case or same court trigger bar discipline referrals as a matter of course? The Arizona Bam pattern (two proceedings, same court, same conduct) and the Tennessee referral to the Board of Professional Responsibility suggest escalation is becoming standard—but the threshold for referral versus continued monetary sanctions remains undefined (→ Arizona Attorney Maren Bam Faces Sanctions for Bogus AI-Generated Quotes in Employment Case, Tennessee Firm Sanctioned $45K for AI-Generated Fake Citations in Malpractice Suit).
• Where does the duty run in capital transactions—to the founder or only the entity? The Simpson Thacher trial will produce a jury finding on this question; the outcome will directly affect how firms scope and document client relationships in fundraising and M&A engagements .
• Does counsel's duty to anticipate regulatory resistance extend to novel business models? The Holland & Knight suit tests whether a firm advising on a consumer-facing fee structure was obligated to flag multi-state AG exposure before it materialized—a duty-scope question with broad implications for fintech, proptech, and platform-economy engagements .

What to watch.

• Whether any state supreme court or federal district adopts Connecticut-style mandatory verification rules with case-ending sanctions—Connecticut's Practice Book rule is the leading candidate for replication .
• Whether the Arizona fee sanction against Maren Bam is imposed and whether the court escalates to a bar referral—the repeat-offender pattern makes this the most visible test of whether monetary sanctions remain the ceiling (→ Arizona Attorney Maren Bam Faces Sanctions for Bogus AI-Generated Quotes in Employment Case).
• Jury verdict in the Simpson Thacher Patriot National trial on the personal-versus-entity duty question—findings on causation and duty scope will reshape engagement-letter drafting in capital transactions .
• Whether any bar ethics opinion adopts or rejects the "negligence for non-use" framing—if adopted, it would create a doctrinal pincer requiring firms to demonstrate both AI competence and AI verification simultaneously (→ Lawyer Faces Potential Negligence Claim for Failing to Use AI in Legal Work).
• Outcome of the Texas TCPA class action against mass tort firms over AI-generated solicitation calls—will define whether existing consumer protection statutes reach AI-driven legal intake workflows .
• Whether malpractice insurers begin conditioning coverage or pricing on documented AI governance protocols, creating a market-driven verification standard independent of bar guidance.