Multi-level Value Alignment in Agentic AI Systems: Survey and Perspectives

TL;DR

This paper surveys multi-level value alignment in agentic AI systems, especially LLM-based multi-agent systems, examining principles, scenarios, methods, and coordination to address societal and systemic risks.

Contribution

It introduces a comprehensive multi-level framework for understanding and evaluating value alignment in agentic AI, integrating principles, applications, and methodologies.

Findings

Structured value principles across macro, meso, micro levels

Categorized application scenarios from general to specific

Mapped alignment methods to a tiered evaluation framework

Abstract

The ongoing evolution of AI paradigms has propelled AI research into the agentic AI stage. Consequently, the focus of research has shifted from single agents and simple applications towards multi-agent autonomous decision-making and task collaboration in complex environments. As Large Language Models (LLMs) advance, their applications become more diverse and complex, leading to increasing situational and systemic risks. This has brought significant attention to value alignment for agentic AI systems, which aims to ensure that an agent's goals, preferences, and behaviors align with human values and societal norms. Addressing socio-governance demands through a Multi-level Value framework, this study comprehensively reviews value alignment in LLM-based multi-agent systems as the representative archetype of agentic AI systems. Our survey systematically examines three interconnected dimensions: First, value principles are structured via a top-down hierarchy across macro, meso, and micro levels. Second, application scenarios are categorized along a general-to-specific continuum explicitly mirroring these value tiers. Third, value alignment methods and evaluation are mapped to this tiered framework through systematic examination of benchmarking datasets and relevant methodologies. Additionally, we delve into value coordination among multiple agents within agentic AI systems. Finally, we propose several potential research directions in this field.