AI summaryⓘ
The authors propose a system to govern AI agents continuously by controlling how much computing power they get, making this control self-enforcing through compute budgets. Their method involves verified human stakeholders making decisions by contributing a special governance currency, which is then aggregated to decide if the AI gets authorization to use compute resources. The system uses a two-level threshold to turn support into a clear yes-or-no decision, and a safety limit ensures the AI’s compute usage stays within safe boundaries. The authors also discuss challenges like preventing the AI from manipulating the decision makers themselves. Overall, their work models AI governance as a game with human contribution driving compute resource allocation.
mechanism designparticipatory governancecompute budgetauthorizationextensive form gamegovernance currencysafety ceilingself-enforcingcompute licensestakeholder
Authors
Praphul Chandra, Sujit Gujar, Ganesh Ghalme
Abstract
We give a formal mechanism design model for the continuous participatory governance of a deployed AI agent. The mechanism is built on the principle that governance should control an AI agent through resource allocation so as to make authorization self enforcing via compute budgets. The mechanism seeks to establish the Safe AI paradigm that compute is an effective governance lever. We situate our work as a compliance or commons overlay on a deployer. One governance period is an extensive form game in which verified human stakeholders arrive sequentially and contribute, on a provision or a rejection market, in a governance currency that is deliberately distinct from the agents compute. A funding aggregator turns raw contributions into breadth weighted effective supports - a two threshold gate with hysteresis converts net support into a binary authorization that, through a coupling map bounded by an exogenously certified safety ceiling, releases a metered compute budget - realized in hardware as a signed compute license so that the decision is self-enforcing. We characterize the class of agents the mechanism can govern and isolate manipulation of the governing electorate by the governed agent as the central open problem. We also introduce several challenges addressing manipulation of governing electorate by the governed agents.