Why Accurate Legal Messages Still Get Misunderstood: Using AI to Pressure-Test Communication with Officials, Boards, and the Public

Local government lawyers regularly have to explain nuanced legal conclusions to elected officials, boards, department leaders, law-enforcement personnel, and members of the public—audiences who may not share the lawyer’s vocabulary, assumptions, or objectives.

Even when a message is legally accurate, it may be interpreted through a political, operational, or emotional lens. A recommendation may be heard as obstruction, an explanation as evasion, or a warning as taking sides.

In this interactive session, participants will learn a practical process for anticipating those interpretation gaps before a message is delivered. They will see how generative AI can be used as a structured role-playing and pressure-testing tool—not as a source of law or legal advice—to simulate how different stakeholders might receive the same communication.

The session will demonstrate how to prompt AI to challenge rather than simply affirm the user’s assumptions, identify likely unintended inferences, and revise a message without sacrificing legal precision. Examples will be drawn from municipal settings such as communicating with city councils and zoning boards, managing public-comment situations, responding to First Amendment auditors, explaining politically sensitive recommendations, and addressing high-emotion police-related matters. All demonstrations will use hypothetical or appropriately sanitized material.

Participants will leave with a four-step process—clarify, simulate, validate, and refine—and sample prompts they can use to pressure-test written advice, public-facing explanations, presentations, and recommendations before they go out.

Learning objectives

By the end of the session, participants will be able to:

• Identify why legally accurate messages may be interpreted differently by nonlawyer stakeholders.
• Anticipate how role, incentives, political pressure, and emotion may shape what an audience hears.
• Use generative AI to produce alternative stakeholder interpretations while avoiding overly agreeable or sycophantic responses.
• Revise a message to reduce unintended interpretations without sacrificing legal accuracy or confidentiality.

Speaker: Daniel Steinberg