As artificial intelligence becomes increasingly embedded in financial services, concerns are growing about less visible forms of exclusion embedded within algorithmic decision-making. Existing research has largely examined algorithmic bias as isolated decision-level outcomes, with limited attention to how disadvantage accumulates and persists over time. This paper develops the concepts of algorithmic vulnerability and compound algorithmic vulnerability to explain how exclusion emerges within AI-mediated financial service ecosystems.
Drawing on Transformative Service Research, Service Ecosystem Theory and literature on AI and consumer vulnerability, this conceptual study develops a multi-level framework comprising algorithmic structuration, experiential vulnerability and ecosystemic mediation. The framework identifies four generative mechanisms: algorithmic redlining, behavioural misclassification, dark nudging and exploitation and identity and privacy risks.
Introduces compound algorithmic vulnerability as an ecosystem-level condition in which multiple vulnerabilities interact, accumulate, and reinforce one another through recursive feedback processes. The framework demonstrates that exclusion in AI-mediated financial services is not simply the result of isolated bias but emerges from interconnected socio-technical mechanisms operating across systems, consumers, and governance structures.
Highlights the need for fairness auditing, inclusive design, explainability and coordinated governance to mitigate algorithmic exclusion and promote financial inclusion.
Advances theory by introducing compound algorithmic vulnerability as a novel lens for understanding how digital disadvantage accumulates within AI-mediated financial services, shifting attention from isolated algorithmic bias to recursive and systemic patterns of exclusion.
