Governing fiduciary relationships or building up a governance model for trust in AI? : Review of healthcare as a socio-technical system

Abstract

Fiduciary law aims to mitigate the inherent risk of ‘trust’, which helps restore interpersonal trust. It remains to be answered how trust should be governed in an AI-driven socio-technical system where technical and social factors are involved including interpersonal relationships and AI-human interactions. Taking interpersonal trust as the backdrop of analysis, this article seeks answers to this question focusing on healthcare. It firstly draws a conceptual framework regarding 'trust' and investigates its interplay with AI as well as examines how it is governed under the fiduciary law. Subsequently, it upholds a socio-technical system perspective, examining how to enable and sustain trust in an AI-driven socio-technical system. A governance model is then developed to elicit ‘intrinsic’, ‘dynamic’ and ‘ethical’ values of trust attributed to various elements under a tri-partite framework. It is recognised that findings of the literature as to trust, its trajectory and implications can be implemented within the proposed framework. Furthermore, it brings novelty by re-conceptualising the elements of 'trust' and associated values, marking distinction to its interpersonal roots and fiduciary relationships. It is considered this governance model, by upholding a holistic viewpoint, provides a generalisable framework that can construct, maintain and restore trust in AI-driven socio-technical systems.