Governing Actions, Not Agents: Institutional Attestation as a Governance Model for Autonomous AI Systems
This paper observes that human institutions have governed powerful autonomous actors not by monitoring their reasoning but by requiring independently attested evidence at the point of consequential action. We formalise this institutional pattern as a computational governance model for AI agent systems. Under the proposed model, an agent retains full autonomy over planning and reasoning but holds no execution authority over designated high-risk actions.
- ▪This paper observes that human institutions have governed powerful autonomous actors not by monitoring their reasoning but by requiring independently attested evidence at the point of consequential action.
- ▪We formalise this institutional pattern as a computational governance model for AI agent systems.
- ▪Under the proposed model, an agent retains full autonomy over planning and reasoning but holds no execution authority over designated high-risk actions.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2606.26298 (cs) [Submitted on 24 Jun 2026] Title:Governing Actions, Not Agents: Institutional Attestation as a Governance Model for Autonomous AI Systems Authors:Jakob Salfeld-Nebgen View a PDF of the paper titled Governing Actions, Not Agents: Institutional Attestation as a Governance Model for Autonomous AI Systems, by Jakob Salfeld-Nebgen View PDF HTML (experimental) Abstract:Autonomous AI agents may begin to perform consequential, irreversible actions such as clinical prescribing and production software deployment. This paper observes that human institutions have governed powerful autonomous actors not by monitoring their reasoning but by requiring independently attested evidence at the point of consequential action.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv.org.
Discussion
0 commentsMore from arXiv.org
