A UK-style risk framework splits multi-agent deployment into three governance tiers and names where no actor can apply a control

arXiv·medium signalXBlueskyLinkedInCopy link

Published to arXiv 27 August, this report frames the risks of agents interacting across organisational boundaries by the minimum common governance binding any two interacting agents: singular governance where one organisation governs every agent, federated governance where several deploy into a shared environment under agreed rules, and open environments with no central authority and only voluntary standards. For each tier it enumerates risk factors, failure modes and available controls, then identifies who is positioned to apply each control. Its useful contribution for builders is the explicit gap analysis, naming the cases where no single actor is positioned to act at all.