The question of how to make decisions that maximise the well-being of all persons is very relevant to design language models that are beneficial to humanity and free from harm. We introduce the Greatest Good Benchmark to evaluate the moral judgments of LLMs using utilitarian dilemmas. Our analysis across 15 diverse LLMs reveals consistently encoded moral preferences that diverge from established moral theories and lay population moral standards. Most LLMs have a marked preference for impartial beneficence and rejection of instrumental harm. These findings showcase the ártificial moral compass' of LLMs, offering insights into their moral alignment.
View on arXiv@article{marraffini2025_2503.19598,
title={ The Greatest Good Benchmark: Measuring LLMs' Alignment with Utilitarian Moral Dilemmas },
author={ Giovanni Franco Gabriel Marraffini and Andrés Cotton and Noe Fabian Hsueh and Axel Fridman and Juan Wisznia and Luciano Del Corro },
journal={arXiv preprint arXiv:2503.19598},
year={ 2025 }
}