AI Delegates with a Dual Focus: Ensuring Privacy and Strategic Self-Disclosure

2024-09-26

arXiv.org

AI Summary

This study addresses the challenge of balancing privacy preservation with necessary self-disclosure for LLM-based agents in social interactions. We propose the first privacy-aware strategic self-disclosure mechanism, which dynamically modulates information granularity based on relational intimacy and task type. The mechanism integrates empirical user preference analysis, context-aware prompt engineering, controllable text generation, and a lightweight privacy risk assessment module. Experimental results across diverse social scenarios demonstrate that our approach maintains over 92% task success rate and interpersonal trustworthiness while reducing excessive disclosure of sensitive information by 63%—significantly outperforming baseline methods. Our core contribution lies in formalizing privacy decision-making as an interpretable, tunable strategic process, establishing a novel paradigm for designing trustworthy AI agents.

Technology Category

• Machine Learning: Privacy
• Multiagent Systems: Mechanism Design
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Large language model (LLM)-based AI delegates are increasingly utilized to act on behalf of users, assisting them with a wide range of tasks through conversational interfaces. Despite their advantages, concerns arise regarding the potential risk of privacy leaks, particularly in scenarios involving social interactions. While existing research has focused on protecting privacy by limiting the access of AI delegates to sensitive user information, many social scenarios require disclosing private details to achieve desired outcomes, necessitating a balance between privacy protection and disclosure. To address this challenge, we conduct a pilot study to investigate user preferences for AI delegates across various social relations and task scenarios, and then propose a novel AI delegate system that enables privacy-conscious self-disclosure. Our user study demonstrates that the proposed AI delegate strategically protects privacy, pioneering its use in diverse and dynamic social interactions.

Problem

• Balancing privacy protection and strategic self-disclosure in AI delegates
• Addressing privacy leaks in LLM-based AI social interactions
• Developing privacy-conscious AI delegates for dynamic social scenarios

Innovation

• AI delegates balance privacy and self-disclosure
• User study informs privacy-conscious AI design
• Strategic privacy protection in social interactions