Q-COTRUST: Economic and Governance Boundaries of Quantum Coordination in Distributed Human–AI Decisions

Distributed Human–AI decisions increasingly allocate scientific verification, public services, and autonomous actions across agents that hold private information, yet communication, verification, and review consume scarce economic and computational resources. Quantum correlations can expand zero-message coordination, but computational capability has economic significance only if it beats feasible substitutes after implementation and oversight are priced. Existing quantum multi-agent studies and benchmark frameworks rarely identify these margins within one matched evaluation. We introduce Q-COTRUST, a role-separated diagnostic benchmark testing whether quantum-assisted coordination retains modeled adoption value—the utility difference between configurations net of specified charges—against feasible classical substitutes, finite quantum resources, and technical controls. Our synthetic two-site scenario routes valuable verification tasks to simulated human or AI workers under private bits, delayed reliability audits, limited messaging, scarce pairs, and capacity-limited review. Q-COTRUST combines an eight-cell structural grid over information, payoff, and communication with a separate eight-regime study of an ideal policy ($A$), pair infrastructure ($I$), and technical controls ($T$), evaluated over 30 paired seeds and an exact labelled-$K_3$ rendezvous check. In the selected local, joint-payoff, no-message cell, the full regime exceeds an ungated classical backbone by $+0.076,[+0.070,+0.083]$ scenario utility per task; yet the governed classical mean is higher ($0.392$ versus $0.389$), the paired 111–001 interval is unavailable, finite realization changes utility by $-0.024,[-0.032,-0.016]$ relative to ideal $A+T$ at the tested prices, and a timely classical bit exceeds the ideal no-message Bell policy ($2/3$ versus $7/12$). These findings recast quantum advantage as an economic–computational identification problem and yield a staged rule: diagnose the bottleneck, benchmark the strongest timely classical alternative, then price realization and control. Because the evidence is synthetic and two-site, it cannot support procurement or deployment claims; that limit defines a three-stage future agenda—artifact validation, price–capacity frontiers, and measured network and participatory field studies—for testing transfer and identifying who bears coordination errors and costs.