Grading open text legal answers with trustworthy rubric and human review
SoK: Grading Open-Text Legal Answers, from Law School Exams to Trustworthy Rubric Judges
Computation and Language
Summary
Grading law school exams is hard because answers are long and open-ended, and human graders often disagree. The authors looked at ways to judge these answers, especially using large language models (LLMs) with detailed scoring guides called rubrics. They found that creating clear rubrics for each question helps computers grade more like humans, but human oversight is still essential to catch alternative correct answers and mistakes. Their approach makes grading more reliable and transparent by combining automated scoring with human review.
What this means in practice
- •For legal education administrators: Improve exam scoring by integrating rubric-based AI grading with human review to make grading faster and more consistent.
- •For legal ai developers: Create grading AI systems that include instance-specific rubrics, evidence checks, and calibration against human graders for trustworthy evaluations.
A survey. It maps existing work.
Authors
Sebastian Nagl
Abstract
Law students are examined through open-text answers: case solutions, essays and memos that apply the law to a set of facts. Such answers are expensive to grade, human graders disagree widely on them, and large language models (LLMs) now offer to grade them at scale, both in teaching and in the benchmarks that test legal AI. This paper systematizes how open-text legal answers can be evaluated. We derive eight desiderata from educational measurement and from the law that governs exams, among them validity, reliability, reasons, credit for defensible alternative solutions, robustness to gaming and human oversight. We then compare eight evaluation paradigms against them, from unstructured human grading to LLM judges with instance-specific rubrics and pairwise comparison. Three findings stand out. A rubric written for the task is the strongest lever for agreement between an LLM judge and human graders, but the reported figures answer different questions and rarely come with the human agreement they must be read against. The central validation gap lies where examination law is most demanding: defensible alternative solutions, follow-on errors, and the examiner's correction of model judgments. And rubric-centred grading is the most auditable family, because it makes the standard, its application and the final decision checkable. From this we derive the trustworthy rubric judge: an instance rubric that credits arguments and alternative paths,evidence checked in code, probes, repeated runs, calibration against independent human grades,logging, and an examiner who reads every script. We state what each safeguard rests on, what remains to be tested, and the uses the evidence supports today.