Plurals: A System for Guiding LLMs Via Simulated Social Ensembles

TL;DR

Plurals is a system that uses simulated social ensembles of language model agents with diverse viewpoints to facilitate pluralistic AI deliberation, aiming to address biases and promote balanced perspectives.

Contribution

Introduces Plurals, a novel Python library that enables customizable, multi-agent deliberation with social structures, integrating real-world data and democratic templates for diverse AI outputs.

Findings

Simulated focus groups aligned with target audience opinions in 75% of trials

Demonstrated fidelity to theoretical constructs through six case studies

System effectively generates diverse viewpoints using social ensemble simulations

Abstract

Recent debates raised concerns that language models may favor certain viewpoints. But what if the solution is not to aim for a 'view from nowhere' but rather to leverage different viewpoints? We introduce Plurals, a system and Python library for pluralistic AI deliberation. Plurals consists of Agents (LLMs, optionally with personas) which deliberate within customizable Structures, with Moderators overseeing deliberation. Plurals is a generator of simulated social ensembles. Plurals integrates with government datasets to create nationally representative personas, includes deliberation templates inspired by deliberative democracy, and allows users to customize both information-sharing structures and deliberation behavior within Structures. Six case studies demonstrate fidelity to theoretical constructs and efficacy. Three randomized experiments show simulated focus groups produced output resonant with an online sample of the relevant audiences (chosen over zero-shot generation in 75% of trials). Plurals is both a paradigm and a concrete system for pluralistic AI. The Plurals library is available at https://github.com/josh-ashkinaze/plurals and will be continually updated.