Answers are grounded in this paper and cite their sources.
Human oversight is widely required for autonomous and artificial-intelligence systems, yet assigning a human role does not establish that the person can influence a hazardous trajectory. This study develops a bounded Safety-Control Governance Framework for assessing whether documentary human-control claims are operationally inspectable. Drawing on systems-theoretic safety, supervisory control, situation awareness, resilience engineering, meaningful human control, and AI governance, the framework distinguishes seven operational functions—authority allocation, feedback and observability, situational comprehension, intervention timing, safe-state objective, escalation, and recovery—and two cross-cutting governance functions: accountability assignment and documentation or assurance. The framework was applied to a purposive corpus of 24 public documents spanning autonomous mobility, maritime autonomy, drones and emergency operations, and high-risk AI or command support, producing 216 document–function observations. Authority and accountability were Explicit in 20 of 24 documents and documentation or assurance in 23 of 24. Explicit specification was less frequent for intervention timing (8 of 22 known cases), safe-state objectives (7 of 23), escalation (10 of 24), and recovery (6 of 24). Documentary families supplied complementary evidence, while configuration analysis located distinct discontinuities, including explicit authority while feedback and comprehension were not both Explicit in ten documents. Most degradation-pattern evidence was anticipatory or prescriptive rather than event-based. These numerical profiles are conditional on the final coding rules and source selection. The framework provides an evidence-linked protocol for tracing documentary control gaps; it does not assess operational performance.
No takes yet. Share an insight, caveat, or question.
Karim Hardy (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: