Who Gets the Kidney? Human-AI Alignment, Indecision, and Moral Values

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Dickerson, John P., Hosseini, Hadi, Khanna, Samarth, Pierce, Leona
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914487260413952
author Dickerson, John P.
Hosseini, Hadi
Khanna, Samarth
Pierce, Leona
author_facet Dickerson, John P.
Hosseini, Hadi
Khanna, Samarth
Pierce, Leona
contents The rapid integration of Large Language Models (LLMs) in high-stakes decision-making -- such as allocating scarce resources like donor organs -- raises critical questions about their alignment with human moral values. We systematically evaluate the behavior of several prominent LLMs against human preferences in kidney allocation scenarios and show that LLMs: i) exhibit stark deviations from human values in prioritizing various attributes, and ii) in contrast to humans, LLMs rarely express indecision, opting for deterministic decisions even when alternative indecision mechanisms (e.g., coin flipping) are provided. Nonetheless, we show that low-rank supervised fine-tuning with few samples is often effective in improving both decision consistency and calibrating indecision modeling. These findings illustrate the necessity of explicit alignment strategies for LLMs in moral/ethical domains.
format Preprint
id arxiv_https___arxiv_org_abs_2506_00079
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Who Gets the Kidney? Human-AI Alignment, Indecision, and Moral Values
Dickerson, John P.
Hosseini, Hadi
Khanna, Samarth
Pierce, Leona
Computers and Society
Artificial Intelligence
Machine Learning
I.2.1; I.2.7; I.2.11
The rapid integration of Large Language Models (LLMs) in high-stakes decision-making -- such as allocating scarce resources like donor organs -- raises critical questions about their alignment with human moral values. We systematically evaluate the behavior of several prominent LLMs against human preferences in kidney allocation scenarios and show that LLMs: i) exhibit stark deviations from human values in prioritizing various attributes, and ii) in contrast to humans, LLMs rarely express indecision, opting for deterministic decisions even when alternative indecision mechanisms (e.g., coin flipping) are provided. Nonetheless, we show that low-rank supervised fine-tuning with few samples is often effective in improving both decision consistency and calibrating indecision modeling. These findings illustrate the necessity of explicit alignment strategies for LLMs in moral/ethical domains.
title Who Gets the Kidney? Human-AI Alignment, Indecision, and Moral Values
topic Computers and Society
Artificial Intelligence
Machine Learning
I.2.1; I.2.7; I.2.11
url https://arxiv.org/abs/2506.00079