Rules 5.1 and 5.3 of the ABA Model Rules of Professional Conduct address partner and supervisory responsibility. They establish that partners are responsible for ensuring that subordinate attorneys and non-attorney staff comply with the Rules of Professional Conduct, and that this responsibility requires affirmative action — not just an absence of knowing misconduct.
In the context of AI, these rules create a specific and underappreciated risk for managing partners at small firms.
The Supervisory Responsibility Framework
Rule 5.1(a) requires partners to make reasonable efforts to ensure that the firm has measures in place giving reasonable assurance that all attorneys in the firm conform to the Rules of Professional Conduct. Rule 5.1(b) applies the same standard to supervising attorneys with respect to attorneys they directly supervise. Rule 5.3 extends these obligations to non-lawyer staff.
The operative phrase is "make reasonable efforts to ensure that the firm has measures in place." This is not satisfied by telling attorneys not to misuse AI. It requires documented systems.
Why AI Creates Specific Supervisory Exposure
The traditional supervisory responsibility framework was built around human work product. An associate drafts a brief, a partner reviews it, errors are caught. The supervisory chain creates both accountability and error correction.
AI-assisted work product disrupts this chain in a specific way: it produces output that looks finished. An AI-generated brief with fabricated citations does not look like a rough draft. It looks like a polished document from a competent associate. Unless the supervising partner independently verifies the citations — or has a documented protocol requiring someone to do so — the fabrication can pass through review undetected.
This is not a hypothetical failure mode. It is the documented mechanism behind a substantial portion of the Rule 3.3 sanctions in KeaneAdvisors.AI's tracker. The supervising partner reviewed the brief. The supervising partner missed the fabricated citations. The court did not.
What "Reasonable Efforts" Requires
Under current case law and bar guidance, reasonable efforts to supervise AI use at a small firm require, at minimum:
An explicit AI use policy that attorneys and staff are required to read and acknowledge. The policy must address the firm's expectations for AI tool approval, client data handling, and output verification.
A defined verification protocol for AI-generated work product. The protocol must specify what types of AI output require independent verification, who is responsible for that verification, and how it is documented.
Regular supervision check-ins that include AI use as a topic. Supervisors should know what tools supervised attorneys and staff are using and how.
A record of compliance. The firm should maintain documentation showing that these measures exist, that they are communicated, and that they are enforced. In a disciplinary proceeding, this documentation is the difference between "we had reasonable measures in place" and "we hope the bar takes our word for it."
The Personal Liability Dimension
Partners at small firms are often exposed in a way that partners at large firms are not, because at a small firm, the partner is also the managing partner, and is personally accountable for the firm's professional conduct measures.
When the ABA and state disciplinary authorities are evaluating a Rule 5.1 or 5.3 matter, they are looking at what the responsible partner did to establish governance. A managing partner who can demonstrate documented, communicated, and enforced AI governance measures is in a fundamentally different position than one who cannot.
The Phase 0 Assessment's Supervision & Accountability domain scores exactly this — whether the firm has defined roles, documented protocols, and created the record that demonstrates responsible AI supervision. For managing partners, the score in this domain is the most direct indicator of their personal professional exposure.