CLIENT ALERT: LEGAL PRACTICE & AI ETHICS

Managing Ethical Risks, Liability, and Risk Mitigation in Attorney Use of Generative AI

TO: Law Firm Partners, General Counsel, Executive Committees, and Managing Attorneys

FROM: Legal Ethics & Risk Management Practice Group

DATE: August 2026

SUBJECT: Ethical Duties, Malpractice Exposure, and Standardized Engagement Safeguards for Generative AI

1. EXECUTIVE SUMMARY & INDUSTRY OVERVIEW

Generative Artificial Intelligence (GAI) tools—ranging from sophisticated legal research engines to automated document drafting suites—have rapidly transitioned from novel experimental software to essential legal infrastructure. While GAI offers transformative gains in efficiency, data organization, and initial work product generation, its undisclosed or unsupervised deployment creates severe professional liability and regulatory exposure. Courts, state bars, and federal tribunals have aggressively sanctioned attorneys for unverified court filings containing hallucinated case citations, unauthorized exposure of client confidential data, and non-transparent billing practices. To maintain compliance with ABA Formal Opinion 512 and evolving jurisdiction-specific ethical rules, law firms must establish strict risk-mitigation protocols and integrate comprehensive GAI disclosures into client engagement documentation.

2. THE ETHICAL AND LIABILITY TRIAD

Attorneys deploying GAI systems face critical ethical exposure across five core provisions of the ABA Model Rules of Professional Conduct, categorized into three principal compliance pillars.

Regarding the compliance framework for Rules 1.1 and 3.3 covering Competence and Candor, attorneys must perform a mandatory independent check of all AI outputs, adhere to a strict prohibition against unverified precedent, and accept absolute liability for false claims or hallucinations presented to courts. GAI models operate probabilistically, creating inherent risks of hallucinated legal precedents, inaccurate statutory interpretations, or missing negative authority. Rule 1.1 obligates practitioners to understand the benefits and risks associated with technologies used in client matters, while Rule 3.3 prohibits making false statements of law or fact to a tribunal. Submitting unverified GAI outputs directly breaches both duties, exposing counsel to monetary sanctions, show-cause orders, and formal disciplinary proceedings.

Regarding Rule 1.6 concerning Privacy and Confidentiality, practices require a strict prohibition on open public model inputs, zero-data-retention and non-training guarantees, and protection against privilege waiver and data breaches. Submitting client facts, trade secrets, financial records, or litigation strategy into consumer-grade, open-loop GAI applications risks waiving attorney-client privilege and violating Rule 1.6. Attorneys must ensure third-party vendors guarantee end-to-end encryption, maintain zero-data-retention policies, and explicitly prohibit the use of prompt inputs for base model training or external human review.

Regarding Rule 1.5 concerning Transparent Billing, attorneys must bill only actual time spent prompting and reviewing, refrain from billing hypothetical time saved, and ensure standard tech overhead is absorbed by the firm. While GAI drastically reduces drafting and research timelines, ABA Formal Opinion 512 mandates that hourly-billing attorneys charge exclusively for actual time expended supervising, prompting, and verifying outputs. Firms are strictly prohibited from billing phantom hours—such as billing 4 hours for a task completed via AI in 20 minutes—or passing routine software licensing fees onto clients as direct disbursements without express prior authorization.

3. CONTRACTUAL RISK MITIGATION: COMPREHENSIVE ENGAGEMENT CLAUSE

To ensure full compliance and maintain client trust, law firms must implement clear, standardized disclosure clauses within retainer agreements and engagement letters using the sample engagement clause for the use of advanced legal technologies and generative AI.

Under paragraph (a) addressing Authorized Scope and Human Supervision, to provide efficient, cost-effective, and high-quality legal representation, the Firm may utilize secure, enterprise-grade software tools, including artificial intelligence and generative AI platforms ("GAI Tools"), to assist with legal research, document analysis, data organization, and initial draft preparation. Client acknowledges and agrees that all GAI Tools function strictly as supplementary administrative and research aids. Every output, analysis, legal argument, or document draft generated with the assistance of GAI Tools shall be subjected to direct, independent human review, verification, and legal assessment by a licensed attorney of the Firm prior to final delivery, reliance, or submission to any court or third party. The Firm maintains ultimate professional responsibility for all final work product.

Under paragraph (b) addressing Protection of Confidentiality and Data Security, the Firm shall not input any of Client's confidential information, trade secrets, protected health data, or privileged communications into any public, unencrypted, or open-access consumer AI application. The Firm warrants that any GAI Tool deployed in connection with Client's representation operates within a secure enterprise environment featuring end-to-end encryption, strict access controls, and explicit vendor commitments that Client data will neither be retained for third-party training purposes nor utilized to improve public AI models.

Under paragraph (c) addressing Transparent Billing and Cost Structure, in accordance with ethical standards governing professional fees, Client will be billed solely for the actual time expended by the Firm's attorneys and staff in operating, prompting, reviewing, modifying, and validating work assisted by GAI Tools. The Firm will not charge premiums, artificial hourly markups, or hypothetical time-saved values for efficiency gains realized through technology. Direct third-party software usage fees for specialized, matter-specific GAI platforms will only be passed through as disbursements if explicitly approved in writing by Client in advance; standard enterprise GAI software subscriptions shall be absorbed as internal Firm overhead.

4. ACTION PLAN FOR LAW FIRM LEADERSHIP

First, regarding Audit Firm Software & Eliminate Shadow AI, conduct an immediate audit of all software used by fee earners, blocking consumer-grade, unencrypted LLM platforms on all firm devices.

Second, regarding Mandate Human-in-the-Loop (HITL) Verification, enforce a strict policy requiring attorneys to verify every case citation, quotation, and statutory reference against primary legal databases prior to filing.

Third, regarding Update Onboarding & Retainer Protocols, incorporate the standardized GAI engagement clause into all upcoming client agreements and outside counsel guidelines.

Fourth, regarding Vet Vendor Security & Data Terms, obtain executed Business Associate Agreements (BAAs) and enterprise vendor commitments confirming zero data retention and non-training compliance.

Note: For tailored assistance updating firm-wide AI governance policies, outside counsel guidelines, or fee schedules, please contact the Firm's Professional Responsibility Practice Group.