Filtered Humanity and Algorithmic Access Part II: Duty of Care and Ethical Risk in AI-Mediated Triage Systems

Abstract

This paper examines the ethical and emerging legal implications of AI-mediated service triage systems that increasingly function as gatekeepers to human assistance across customer service, healthcare administration, and public-sector interfaces. These systems are commonly introduced to improve efficiency, manage demand, and reduce operational cost. Yet in doing so, they quietly restructure access to care by determining which individuals are escalated to human attention and which are filtered out through automated classification processes. Drawing on organisational ethics, governance theory, and duty-of-care principles, the paper argues that algorithmic triage represents a form of hybrid automation in which moral responsibility is redistributed while remaining formally unacknowledged. Individuals seeking help are required to translate distress, urgency, or complexity into machine-interpretable signals, while organisations retain plausible deniability for exclusionary outcomes produced by system design rather than human judgement. Using a duty-architecture framework, the paper identifies foreseeable harms arising from silence induction, dignity erosion, emotional displacement, and inequitable access to assistance. It contends that existing legal obligations relating to reasonable care, good faith, and psychological wellbeing are already strained by these systems, even in the absence of AI-specific regulation. The paper concludes that AI-mediated triage constitutes a significant ethical governance risk, not because such systems fail, but because they succeed quietly in narrowing humanity under the banner of optimisation. This paper builds on earlier analysis of hybrid automation by examining how access to human care becomes algorithmically mediated.

Ähnliche Arbeiten