Trust as a cybernetic medium in digital financial inclusion: a multi-level framework

Purpose This article develops a multi-level cybernetic framework conceptualising trust as the primary complexity-reducing medium in digital financial inclusion (DFI) systems, integrating Luhmann's social systems theory, Beer's viable system model (VSM) and the algorithmic-trust accounts of Esposito and Coeckelbergh. Design/methodology/approach Following Jaakkola's (2020) typology, this is a theory-synthesis article. Each source tradition's role is tiered: core ontology (Luhmann), core structure (Beer), algorithmic bridge (Esposito and Coeckelbergh), micro–macro connector (Sessions and institutionalisation theory) and empirical domain (DFI scholarship). Five theoretically derived propositions are developed. Findings DFI is modelled as an operationally semi-autonomous sub-system: self-referentially closed over its own recursive decision communications while remaining programmatically subordinate to the economy's payment code. Within it, trust attenuates environmental variety across three modalities – interpersonal, institutional and algorithmic – mapped onto Beer's five systems via explicit derivation rules that constitute a preferred, rule-governed mapping. Second-order reflexivity explains how trust deficits destabilise inclusion pathways and how trust repair stabilises the ecosystem under turbulence. Practical implications The mapping locates systemic vulnerabilities, the S2 channel specifies how actors at different levels mesh and the framework supports maturity-stage-differentiated regulation and gender-responsive trust architecture (SDG 5). Originality/value The article (1) clarifies the code/programme status of inclusion/exclusion, (2) maps tri-modal trust onto VSM levels through explicit derivation rules, (3) extends Luhmann to algorithmic actors and (4) establishes second-order reflexivity – a recursively interacting, five-component capacity – as an independent determinant of trust resilience.