# From Runtime Records to Legal Findings: An Evidentiary-Adequacy Criterion for Agentic AI Oversight

arXiv: https://arxiv.org/abs/2607.00941 · HTML: https://mlsanity.com/paper/2607.00941/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use.
- Builds on 4 of its 23 references that are in the map.
- First posted on arXiv 1 Jul 2026.

## Placed in

- LLM agents and tool use / [Agent threats, guardrails and governance](https://mlsanity.com/agents/agent-threats-guardrails-and-governance/)

## Builds on (most cited first, 4 of 4)

- [AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents](https://mlsanity.com/paper/2503.18666/) (2025)
- [MI9: An Integrated Runtime Governance Framework for Agentic AI](https://mlsanity.com/paper/2508.03858/) (2025)
- [Formalising Human-in-the-Loop: Computational Reductions, Failure Modes, and Legal-Moral Responsibility](https://mlsanity.com/paper/2505.10426/) (2025)
- [A "good regulator theorem" for embodied agents](https://mlsanity.com/paper/2508.06326/) (2025)

## Built on by (most cited first, 2 of 2)

- [The CASE Framework: A Multi-Disciplinary Control Architecture for Governing Enterprise Agentic AI](https://mlsanity.com/paper/2608.10153/) (2026)
- [The Anatomy of a Blind Spot: What an LLM Judge Must Be Shown, and How Much](https://mlsanity.com/paper/2610.15938/) (2026)

## Abstract (arXiv)

Agentic AI systems generate runtime records, logs, traces, and audit artefacts, but the existence or integrity of such records does not by itself establish that legally operative oversight findings can be recovered from them. This technical report defines an evidentiary-adequacy criterion for a bounded class of determinations: binary findings of fact about specific events and their relations, such as whether protected data crossed a boundary, whether a human could intervene, whether an information barrier held, or whether delegated authority was valid at the moment of use. The criterion states that a runtime record can answer such a determination only if it carries both a typing that maps recorded events to the legally operative category and the relation, such as provenance, authority, derivation, or temporal validity, on which the determination's truth depends. The claim is one of necessity, not sufficiency. The report instantiates the criterion against selected EU AI Act oversight obligations and explains why tamper-proof logs, generic process frameworks, and provenance structures alone cannot establish the relevant findings. It further relates the argument to requisite variety, the Good Regulator Theorem, and the trace-versus-hyperproperty boundary of runtime verification. Companion materials and the experiment protocol are archived on Zenodo.