Social-RAG: Retrieving from Group Interactions to Socially Ground AI Generation

TL;DR

Social-RAG is a method that enhances AI's social appropriateness in group interactions by retrieving and utilizing prior social signals, demonstrated through a system that improved paper recommendations in research groups.

Contribution

The paper introduces Social-RAG, a novel workflow for grounding AI in social group norms by leveraging prior interactions and social signals, implemented in the PaperPing system.

Findings

PaperPing posted relevant messages without disrupting social practices.

The system fostered group common ground over three months.

Social signals improved the relevance and social alignment of AI messages.

Abstract

AI agents are increasingly tasked with making proactive suggestions in online spaces where groups collaborate, yet risk being unhelpful or even annoying if they fail to match group preferences or behave in socially inappropriate ways. Fortunately, group spaces have a rich history of prior interactions and affordances for social feedback that can support grounding an agent's generations to a group's interests and norms. We present Social-RAG, a workflow for socially grounding agents that retrieves context from prior group interactions, selects relevant social signals, and feeds them into a language model to generate messages in a socially aligned manner. We implement this in \textsc{PaperPing}, a system for posting paper recommendations in group chat, leveraging social signals determined from formative studies with 39 researchers. From a three-month deployment in 18 channels reaching 500+ researchers, we observed PaperPing posted relevant messages in groups without disrupting their existing social practices, fostering group common ground.