Measuring the Groundedness of Legal Question-Answering Systems

TL;DR

This paper develops and evaluates methods to assess whether AI-generated legal responses are properly grounded in source material, aiming to improve trustworthiness in high-stakes legal applications.

Contribution

It introduces a comprehensive benchmark and novel evaluation techniques for groundedness in legal question-answering systems, including a new grounding classification corpus.

Findings

Best method achieved macro-F1 score of 0.8

Evaluation of latency for real-world applicability

Demonstrated potential for improving AI trustworthiness in legal settings

Abstract

In high-stakes domains like legal question-answering, the accuracy and trustworthiness of generative AI systems are of paramount importance. This work presents a comprehensive benchmark of various methods to assess the groundedness of AI-generated responses, aiming to significantly enhance their reliability. Our experiments include similarity-based metrics and natural language inference models to evaluate whether responses are well-founded in the given contexts. We also explore different prompting strategies for large language models to improve the detection of ungrounded responses. We validated the effectiveness of these methods using a newly created grounding classification corpus, designed specifically for legal queries and corresponding responses from retrieval-augmented prompting, focusing on their alignment with source material. Our results indicate potential in groundedness classification of generated responses, with the best method achieving a macro-F1 score of 0.8. Additionally, we evaluated the methods in terms of their latency to determine their suitability for real-world applications, as this step typically follows the generation process. This capability is essential for processes that may trigger additional manual verification or automated response regeneration. In summary, this study demonstrates the potential of various detection methods to improve the trustworthiness of generative AI in legal settings.