Trust and explainability in A/IS-mediated healthcare: Operationalizing the therapeutic alliance in a distributed system

Abstract

Healthcare systems are under-resourced. Autonomous and intelligent systems (A/IS) bring the promise of providing methods that can supplement, or replace, healthcare professionals (HCPs). However, healthcare is most effect when patients and HCPs develop a therapeutic alliance (i.e., a trusting relationships) that might not be replicable in A/IS-human interaction. A framework is provided to operationalize the therapeutic alliance in A/IS-mediate healthcare. Using models of persuasive communication as a guide, it assumes that treatment compliance requires understanding discrepancies between the disease models of HCPs and patients in a transactive memory network. The paper also highlights the ethical issues that arise from the creation, or failures of, this process.