Being Considerate as a Pathway Towards Pluralistic Alignment for Agentic AI

Parand A. Alamdari, Toryn Q. Klassen, Rodrigo Toro Icarte, Sheila A. McIlraith
11/15/2024
cs.AIcs.CYcs.LG

No confident code implementation found for this paper.

Abstract

Pluralistic alignment is concerned with ensuring that an AI system's objectives and behaviors are in harmony with the diversity of human values and perspectives. In this paper we study the notion of pluralistic alignment in the context of agentic AI, and in particular in the context of an agent that is trying to learn a policy in a manner that is mindful of the values and perspective of others in the environment. To this end, we show how being considerate of the future wellbeing and agency of other (human) agents can promote a form of pluralistic alignment.

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

Exploring Syntropic Frameworks in AI Alignment: A Philosophical Investigation

82% match2025

AI, Pluralism, and (Social) Compensation

82% match2024

Pluralistic Alignment Over Time

82% match2024

Measuring Value Alignment

80% match2023

Designing for Human-Agent Alignment: Understanding what humans want from their agents

80% match2024

Resource Rational Contractualism Should Guide AI Alignment

80% match2025

Cite this paper

@article{alamdari2024being,
  title  = {Being Considerate as a Pathway Towards Pluralistic Alignment for Agentic AI},
  author = {Parand A. Alamdari and Toryn Q. Klassen and Rodrigo Toro Icarte and Sheila A. McIlraith},
  year   = {2024},
  eprint = {2411.10613},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2411.10613v1}
}

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/2411.10613.svg)](https://codeofpaper.com/paper/2411.10613)

Discussion