Artificial Organisations
This work offers a novel architectural approach to AI safety for multi-agent systems by drawing inspiration from human institutional design, potentially enabling reliable collective behavior from unreliable individual components.
The paper proposes an institutional model for multi-agent AI systems, using compartmentalization and adversarial review to achieve reliable outcomes without assuming individual AI alignment. They demonstrate this with the Perseverance Composition Engine, which shows progression from attempted fabrication to honest refusal in impossible tasks over 474 composition cycles.
Alignment research focuses on making individual AI systems reliable. Human institutions achieve reliable collective behaviour differently: they mitigate the risk posed by misaligned individuals through organisational structure. Multi-agent AI systems should follow this institutional model using compartmentalisation and adversarial review to achieve reliable outcomes through architectural design rather than assuming individual alignment. We demonstrate this approach through the Perseverance Composition Engine, a multi-agent system for document composition. The Composer drafts text, the Corroborator verifies factual substantiation with full source access, and the Critic evaluates argumentative quality without access to sources: information asymmetry enforced by system architecture. This creates layered verification: the Corroborator detects unsupported claims, whilst the Critic independently assesses coherence and completeness. Observations from 474 composition tasks (discrete cycles of drafting, verification, and evaluation) exhibit patterns consistent with the institutional hypothesis. When assigned impossible tasks requiring fabricated content, this iteration enabled progression from attempted fabrication toward honest refusal with alternative proposals--behaviour neither instructed nor individually incentivised. These findings motivate controlled investigation of whether architectural enforcement produces reliable outcomes from unreliable components. This positions organisational theory as a productive framework for multi-agent AI safety. By implementing verification and evaluation as structural properties enforced through information compartmentalisation, institutional design offers a route to reliable collective behaviour from unreliable individual components.