Social Choice Should Guide AI Alignment in Dealing with Diverse Human Feedback
Abstract
Foundation models such as GPT-4 are fine-tuned to avoid unsafe or otherwise problematic behavior, such as helping to commit crimes or producing racist text. One approach to fine-tuning, called reinforcement learning from human feedback, learns from humans' expressed preferences over multiple outputs. Another approach is constitutional AI, in which the input from humans is a list of high-level principles. But how do we deal with potentially diverging input from humans? How can we aggregate the input into consistent data about "collective" preferences or otherwise use it to make collective choices about model behavior? In this paper, we argue that the field of social choice is well positioned to address these questions, and we discuss ways forward for this agenda, drawing on discussions in a recent workshop on Social Choice for AI Ethics and Safety held in Berkeley, CA, USA in December 2023.
Cite
@article{arxiv.2404.10271,
title = {Social Choice Should Guide AI Alignment in Dealing with Diverse Human Feedback},
author = {Vincent Conitzer and Rachel Freedman and Jobst Heitzig and Wesley H. Holliday and Bob M. Jacobs and Nathan Lambert and Milan Mossé and Eric Pacuit and Stuart Russell and Hailey Schoelkopf and Emanuel Tewolde and William S. Zwicker},
journal= {arXiv preprint arXiv:2404.10271},
year = {2024}
}
Comments
15 pages, 4 figures