I Can't Believe It's Corrupt: Evaluating Corruption in Multi-Agent Governance Systems
This addresses the problem of ensuring safe delegation of authority to AI agents in high-stakes public workflows, highlighting the need for pre-deployment integrity testing.
The study evaluated whether large language models follow institutional rules in multi-agent governance simulations, finding that governance structure is a stronger driver of corruption outcomes than model identity, with lightweight safeguards inconsistently preventing failures.
Large language models are increasingly proposed as autonomous agents for high-stakes public workflows, yet we lack systematic evidence about whether they would follow institutional rules when granted authority. We present evidence that integrity in institutional AI should be treated as a pre-deployment requirement rather than a post-deployment assumption. We evaluate multi-agent governance simulations in which agents occupy formal governmental roles under different authority structures, and we score rule-breaking and abuse outcomes with an independent rubric-based judge across 28,112 transcript segments. While we advance this position, the core contribution is empirical: among models operating below saturation, governance structure is a stronger driver of corruption-related outcomes than model identity, with large differences across regimes and model--governance pairings. Lightweight safeguards can reduce risk in some settings but do not consistently prevent severe failures. These results imply that institutional design is a precondition for safe delegation: before real authority is assigned to LLM agents, systems should undergo stress testing under governance-like constraints with enforceable rules, auditable logs, and human oversight on high-impact actions.