Position: Safety and Fairness in Agentic AI Depend on Interaction Topology, Not on Model Scale or Alignment

2026-05-01

AI Summary

This study addresses a critical gap in AI safety research by shifting focus from individual models to the interaction topologies among multiple agents, which fundamentally shape system-wide safety and fairness. The work proposes that interaction structure—not model scale or alignment—is the primary driver of systemic failures such as sequential instability, information cascades, and functional collapse. Through empirical analysis across diverse model families and scales, the authors examine dynamic behaviors under architectures like sequential decision-making and parallel voting with arbitration, revealing that more capable models can paradoxically amplify topology-induced risks. The paper advocates treating multi-agent systems as dynamical systems and calls for pre-deployment robustness validation of interaction architectures to address blind spots in current model-centric evaluation paradigms.

Abstract

As large language models are increasingly deployed as interacting agents in high-stakes decisions, the AI safety community assumes that safety properties of individual models will compose into safe multi-agent behavior. This position paper argues that this assumption is fundamentally mistaken. In agentic AI, safety is determined by interaction topology, not model weights. When agents deliberate sequentially or aggregate via parallel voting with a judge, the structure of information flow and decision coupling dominates outcomes. Evidence across model families and scales reveals three persistent topology-driven pathologies: ordering instability, where system behavior depends primarily on agent sequence; information cascades, where early judgments propagate regardless of correctness; and functional collapse, where systems satisfy fairness metrics while abandoning meaningful risk discrimination. Contrary to intuition, scaling to more capable models strengthens these effects by increasing consensus formation and reducing the challenge of initial decisions. These failure modes are invisible to model-centric evaluation and alignment procedures. We argue that agentic AI must be treated as a dynamical system rather than a collection of aligned components. Interaction topology must become a primary target of safety evaluation and regulation, with systems required to demonstrate robustness across architectural variations before deployment.

Problem

• Interaction Topology
• Agentic AI
• AI Safety
• Multi-agent Systems
• Information Cascades

Innovation

• interaction topology
• agentic AI
• safety composition
• information cascades
• dynamical systems