Automating AI Failure Tracking: Semantic Association of Reports in AI Incident Database

TL;DR

This paper introduces an automated semantic retrieval framework for associating new AI failure reports with existing incidents in the AI Incident Database, improving scalability and accuracy over manual methods.

Contribution

It formalizes report association as a ranking task using transformer-based embeddings, demonstrating superior performance and robustness across various configurations.

Findings

Transformer-based embeddings outperform lexical methods.

Combining titles and descriptions improves ranking accuracy.

Retrieval performance improves with larger training datasets.

Abstract

Artificial Intelligence (AI) systems are transforming critical sectors such as healthcare, finance, and transportation, enhancing operational efficiency and decision-making processes. However, their deployment in high-stakes domains has exposed vulnerabilities that can result in significant societal harm. To systematically study and mitigate these risk, initiatives like the AI Incident Database (AIID) have emerged, cataloging over 3,000 real-world AI failure reports. Currently, associating a new report with the appropriate AI Incident relies on manual expert intervention, limiting scalability and delaying the identification of emerging failure patterns. To address this limitation, we propose a retrieval-based framework that automates the association of new reports with existing AI Incidents through semantic similarity modeling. We formalize the task as a ranking problem, where each report-comprising a title and a full textual description-is compared to previously documented AI Incidents based on embedding cosine similarity. Benchmarking traditional lexical methods, cross-encoder architectures, and transformer-based sentence embedding models, we find that the latter consistently achieve superior performance. Our analysis further shows that combining titles and descriptions yields substantial improvements in ranking accuracy compared to using titles alone. Moreover, retrieval performance remains stable across variations in description length, highlighting the robustness of the framework. Finally, we find that retrieval performance consistently improves as the training set expands. Our approach provides a scalable and efficient solution for supporting the maintenance of the AIID.