Large Language Models in Injury Prediction Tools: Simplifying User Interactions and Improving Risk Interpretation.

Abstract

Advances in Large Language Models (LLMs) offer new opportunities to improve accessibility and usability of finite-element (FE) modeling in injury biomechanics. This study presents an LLM-based tool capable of guiding novice users in selecting response surface models trained on FE simulation results and predicting injury outcomes in Behind Armor Blunt Trauma scenarios. Beyond executing predictive tasks, the LLM-based tool communicates complex injury metrics in clear, non-technical language, facilitating broader understanding and adoption of sophisticated modeling frameworks. These findings highlight the potential of integrating LLMs with FE modeling to bridge expertise gaps, enhance interactivity, and support decision-making in injury prediction and other engineering domains.