How to Use ChatGPT to Check If a Case Is Still Good Law

Why good-law checks need more than a yes or no

ChatGPT can help a user ask better questions about a case. It can identify the case, summarize the issue, and turn a vague concern into a checklist: what court decided it, what proposition the user wants to rely on, what later cases cite it, and whether later treatment changes the risk.

But a good-law question is not answered by confidence alone. A case may still be valid for one point and risky for another. It may be followed on one issue, distinguished on different facts, declined by a different court, or overruled in part. Descrybe Legal Engine helps ChatGPT work through that source-checking workflow with primary-law tools, treatment signals, citing authorities, and source text.

What "good law" can mean

When someone asks whether a case is still good law, they may mean several different things. The case might still exist, but that does not mean it is strong support for the exact proposition the user wants.

• Has the case been overruled or abrogated?
• Has a later case limited the holding?
• Have later courts distinguished the case on important facts?
• Is the case binding in the relevant jurisdiction?
• Does the case support the precise proposition being used?
• Is the user relying on a majority holding, dicta, concurrence, dissent, or procedural point?
• Are there newer authorities that are stronger, clearer, or more directly on point?

How treatment signals help

Treatment signals help organize later discussion of a case. They are useful because a single case can be cited by many later decisions, and not every citation has the same meaning.

Common treatment categories in Descrybe include overruled, declined to follow, distinguished, followed, and mentioned. Those categories help organize the review, especially when ChatGPT is helping turn research results into a practical checklist.

How Descrybe helps ChatGPT check treatment

Descrybe Legal Engine lets ChatGPT use Descrybe tools during a legal research conversation. For good-law checks, that means ChatGPT can help move from a case name or citation to source text, treatment review, citing authorities, and related research.

Descrybe also uses authority-ranked case search for concept research. That ranking is meant to surface likely useful or promising authorities and research value notes. It is not a final good-law determination, but it can help ChatGPT start with better candidates before the user reviews source text and treatment.

• Look up the case by citation, name, or partial reference.
• Retrieve case details and source text.
• Review treatment signals for later cases.
• Find citing authorities and inspect how they discuss the case.
• Search for newer authority on the same issue.
• Use returned source text to decide whether the case remains useful for the specific proposition.

A safer ChatGPT workflow for checking good law

The safest workflow asks ChatGPT to organize the question, but asks Descrybe to bring the legal sources into view. The user should avoid stopping at a one-line answer like "yes" or "no."

• Identify the exact case and citation.
• Identify the proposition the user wants to rely on.
• Ask ChatGPT to use Descrybe Legal Engine to look up the case directly.
• Review treatment signals and citing authorities.
• Read the later cases that seem important.
• Ask ChatGPT to summarize only the verified source findings.
• Decide whether the case remains useful for the specific issue, jurisdiction, and procedural context.

Example: checking a case in ChatGPT

Suppose ChatGPT summarizes an argument and identifies a case that seems to support the user's position. The next question should not be whether ChatGPT sounds confident. The next question should be whether the authority still holds up for the exact point.

The user can ask ChatGPT to use Descrybe Legal Engine to look up the citation, retrieve the opinion, check treatment, find citing cases, and search for newer authority on the same issue. If later cases distinguish the authority, ChatGPT can help organize the distinction, but the user still needs to read the source and decide whether the case remains useful.

What this workflow does not replace

This workflow does not replace Shepard's, KeyCite, a lawyer's professional judgment, or the need to read important cases. Descrybe is a legal research system that helps users find, review, and verify legal sources.

Treatment signals and citing-authority lists are starting points for source review. They help ChatGPT and the user organize the work, but they are not a substitute for deciding whether an authority should be used in a filing, memo, client communication, or legal analysis.