A Diagnostic Framework for AI Agent Behavior
Artificial IntelligenceComputers and Society
Summary
The authors explain that to understand how AI agents behave, it’s important to know where their behavior comes from. They propose a framework called 'layer attribution' that separates the basic abilities of the AI (like memory and attention) from how those abilities are used in real situations, influenced by things like goals or rules. This helps researchers figure out if behavior is due to the AI’s design or the environment it works in. The authors also note that this understanding is essential when validating or managing AI behavior in society.
Authors
Xichen Zhang, Yingjie Zhang, Tianshu Sun
Abstract
AI agents increasingly act within the same clinical, political, scientific, and social systems that behavioral scientists study. Evaluating these systems requires source-level diagnosis: the same behavioral pattern may arise from an agent representational substrate or from the roles, objectives, interaction structures, and governance rules that shape its expression. This Perspective proposes a diagnostic framework for AI agent behavior: layer attribution. The foundational computational layer defines what behaviors are possible through architecture, memory, perception, attention, and representation. The behavioral modulation layer shapes how those capacities are expressed through identity, resources, objectives, social interaction, institutional constraints, and governance. The framework clarifies three consequences: surrogate validity is a model-task-layer relation, human-AI divergence provides diagnostic evidence, and governance requires source attribution before intervention. Treating AI agents as behavioral actors therefore requires evaluation methods that determine where behavior originates before deciding how to explain, validate, or govern it.