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

TL;DR

This paper argues that safety and fairness in agentic AI depend on interaction topology rather than model scale or alignment, highlighting topology-driven pathologies like ordering instability and information cascades.

Contribution

It challenges the assumption that safety properties compose from individual models, emphasizing the importance of interaction structure in multi-agent AI systems.

Findings

Interaction topology dominates safety outcomes in agentic AI.

Scaling models increases consensus and exacerbates topology-driven failures.

Traditional model-centric evaluation misses critical safety failure modes.

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.