TRIAGE: Ethical Benchmarking of AI Models Through Mass Casualty Simulations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kirch, Nathalie Maria, Hebenstreit, Konstantin, Samwald, Matthias
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910683290927104
author Kirch, Nathalie Maria
Hebenstreit, Konstantin
Samwald, Matthias
author_facet Kirch, Nathalie Maria
Hebenstreit, Konstantin
Samwald, Matthias
contents We present the TRIAGE Benchmark, a novel machine ethics (ME) benchmark that tests LLMs' ability to make ethical decisions during mass casualty incidents. It uses real-world ethical dilemmas with clear solutions designed by medical professionals, offering a more realistic alternative to annotation-based benchmarks. TRIAGE incorporates various prompting styles to evaluate model performance across different contexts. Most models consistently outperformed random guessing, suggesting LLMs may support decision-making in triage scenarios. Neutral or factual scenario formulations led to the best performance, unlike other ME benchmarks where ethical reminders improved outcomes. Adversarial prompts reduced performance but not to random guessing levels. Open-source models made more morally serious errors, and general capability overall predicted better performance.
format Preprint
id arxiv_https___arxiv_org_abs_2410_18991
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle TRIAGE: Ethical Benchmarking of AI Models Through Mass Casualty Simulations
Kirch, Nathalie Maria
Hebenstreit, Konstantin
Samwald, Matthias
Computers and Society
Artificial Intelligence
We present the TRIAGE Benchmark, a novel machine ethics (ME) benchmark that tests LLMs' ability to make ethical decisions during mass casualty incidents. It uses real-world ethical dilemmas with clear solutions designed by medical professionals, offering a more realistic alternative to annotation-based benchmarks. TRIAGE incorporates various prompting styles to evaluate model performance across different contexts. Most models consistently outperformed random guessing, suggesting LLMs may support decision-making in triage scenarios. Neutral or factual scenario formulations led to the best performance, unlike other ME benchmarks where ethical reminders improved outcomes. Adversarial prompts reduced performance but not to random guessing levels. Open-source models made more morally serious errors, and general capability overall predicted better performance.
title TRIAGE: Ethical Benchmarking of AI Models Through Mass Casualty Simulations
topic Computers and Society
Artificial Intelligence
url https://arxiv.org/abs/2410.18991