Aligning AI With Shared Human Values

Dan Hendrycks, Collin Burns, Steven Basart, Andrew Critch, Jerry Li +2 more
8/5/2020
cs.CYcs.AIcs.CL

Official code: hendrycks/ethics (329 stars, MIT, last commit 2023-04-21).

Abstract

We show how to assess a language model's knowledge of basic concepts of morality. We introduce the ETHICS dataset, a new benchmark that spans concepts in justice, well-being, duties, virtues, and commonsense morality. Models predict widespread moral judgments about diverse text scenarios. This requires connecting physical and social world knowledge to value judgements, a capability that may enable us to steer chatbot outputs or eventually regularize open-ended reinforcement learning agents. With the ETHICS dataset, we find that current language models have a promising but incomplete ability to predict basic human ethical judgements. Our work shows that progress can be made on machine ethics today, and it provides a steppingstone toward AI that is aligned with human values.

View on arXivView PDF

Code Implementations(1)

hendrycks/ethicsOfficial100%
MIT

Aligning AI With Shared Human Values (ICLR 2021)

32947PythonAug 6, 20203 years agoMIT
ai-safetyethical-aigpt-3machine-ethicsml-safety

Similar Papers

Unsupervised Elicitation of Moral Values from Language Models

89% match2026

ClarityEthic: Explainable Moral Judgment Utilizing Contrastive Ethical Insights from Large Language Models

89% match2024

The Morality of Probability: How Implicit Moral Biases in LLMs May Shape the Future of Human-AI Symbiosis

89% match2025

Beyond Ethical Alignment: Evaluating LLMs as Artificial Moral Assistants

89% match1 repo2025

A Data Fusion Framework for Multi-Domain Morality Learning

89% match2023

The Moral Turing Test: Evaluating Human-LLM Alignment in Moral Decision-Making

89% match2024

Cite this paper

@article{hendrycks2020aligning,
  title  = {Aligning AI With Shared Human Values},
  author = {Dan Hendrycks and Collin Burns and Steven Basart and Andrew Critch and Jerry Li and Dawn Song and Jacob Steinhardt},
  year   = {2020},
  eprint = {2008.02275},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2008.02275}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2008.02275.svg)](https://codeofpaper.com/paper/2008.02275)

Discussion