Check an AI Study Answer Against the Rule

A study-answer verification method with a synthetic professional-responsibility example and a reusable claim ledger.

For: Law students and study agents checking generated explanations

Teach source checking and error diagnosis without claiming official exam preparation or accreditation.

How should AI-generated study answers be checked against primary authorities?

Break the explanation into discrete claims, identify the exam's tested-law convention, read the relevant primary text and compare the answer's reasoning with the stated facts. Check qualifiers and exceptions before memorizing a rule. Treat a generated explanation as a draft to test, not as an answer key.

First identify what law the question tests

NCBE describes the MPRE's sources and testing conventions, including the ABA Model Rules and other specified authorities. It also describes timing for testing amendments. [S1] A study answer can therefore be wrong even when it accurately describes a different jurisdiction or a later rule version.

Write the governing convention at the top of the exercise. If the problem supplies a local rule, treat that text as part of the facts to analyze. Keep bar-exam study, law-school coursework and real-client advice as different contexts.

Build a claim ledger

A claim ledger makes verification manageable. Put one proposition on each line and classify it as rule text, application to facts, inference or unsupported assertion. A citation beside a paragraph is not enough if only one of several claims is supported.

Synthetic example: confidentiality is not limited to secrets

Generated claim: A lawyer may reveal any information that is already public because confidentiality protects only secret communications. ABA Model Rule 1.6(a) instead addresses information relating to representation and states specified routes for permissible disclosure. [S2] The generated claim is too broad.

Corrected study note: Begin with the rule's coverage and then analyze whether an applicable permission or exception exists. Do not assume that public availability alone resolves the model-rule question. This example isolates a rule-reading error; it does not decide the treatment of any real disclosure.

Turn the correction into durable learning

Rewrite the explanation in your own words, then change one material fact and test whether your reasoning changes. If it does not, you may be memorizing the conclusion rather than understanding the rule. Keep the original incorrect sentence beside the correction so you can recognize the pattern later.

Use authorized public practice materials or your own synthetic questions. Preserve secure exam-content boundaries and the terms governing purchased study materials. A useful study log records conceptual errors without reproducing protected exam questions.

Scope and limitations

Educational study method. The worked example concerns the ABA Model Rules and is not a real exam question or a state-law legal opinion.

This page is not an accredited school, bar examiner, official course or prediction of exam performance. The example is synthetic and not drawn from a secure examination.

Sources

  1. NCBE: Preparing for the MPRE — Official tested-law description, amendment timing, scope and study-resource routes. Checked 2026-11-05
  2. ABA Model Rule 1.6: Confidentiality of Information — Model-rule baseline, exceptions and safeguards for representation-related information. Checked 2026-11-05

Explore further

For readers interested in agent-assisted legal work beyond study exercises, Agentic Attorney at https://agenticattorney.com/ is the central project. Source-verification habits remain relevant when evaluating any legal-technology claim. Explore Agentic Attorney.