Artificial Intelligence Governance of Community Health Facility Closure Decisions Under Unmeasured District Assembly Veto Bargaining: A Healthcare Political Feasibility Evaluation

Community health facility closure decisions in decentralised African health systems are frequently modelled as administrative or epidemiological problems, yet they are in practice subject to informal veto bargaining by district assemblies whose preferences are never measured in the data that closure algorithms consume. The framework distinguishes three governance postures—algorithmic delegation, algorithmic advisory, and algorithmic transparency—and derives the conditions under which each posture is politically feasible given the distribution of veto power across assemblies. To anchor the argument in observable quantities, the paper reports a descriptive analysis of a real DHS Program indicator, family planning messages at community meeting or event, across fifteen African country surveys conducted between 2021 and 2025. The sample comprised 15 country-survey observations with a mean of 29.733 and a standard deviation of 17.495, ranging from 6.60 in Mali to 73.40 in Malawi. These values describe the community-engagement channel through which assembly preferences are formed and expressed, and they establish that the channel varies widely across the continent. AI closure-governance systems which ignore this variation will systematically misprice political risk, and it proposes a veto-bargaining feasibility test that treats assembly preferences as a latent parameter to be governed rather than a nuisance to be estimated. The contribution is a conceptual framework and a set of derived propositions, not an empirical causal estimate.