Walker, L. (2024). Artificial narrow intelligence-driven diagnostics: impacts, inequities, and policy imperatives in global healthcare [Dissertation, Technische Universität Wien]. reposiTUm. https://doi.org/10.34726/hss.2024.85320
Artificial Intelligence; AI; AI Ethics; Artificial Narrow Intelligence; Health Policy; Health Inequality
en
Abstract:
The greater societal penetration of Artificial Narrow Intelligence (ANI) has highlighted an unfortunate paradox: although emerging technologies have contributed many positive social benefits, there is evidence to suggest they tend to disproportionately benefit high-resource regions compared to their economically disadvantaged counterparts; thus, exacerbating global inequalities. This trend is pervasive across many industries, yet the impact to global health is particularly poignant. To better understand these phenomena, this research paper employs a comparative case study approach to examine the societal, ethical, and policy paradigms in heterogeneous resource allocations. In total, the sample of literature involves over 175 sources of secondary data. Coding and thematic analysis were used to identify gaps and opportunities for policy improvement in intergovernmental organizations (IGOs). Our study resulted in the identification of the following key themes whereby necessary macro-level policies are either 1) absent 2) insufficient 3) inconsistently-applied 4) misattributed and / or 5) lacking efficacy in real-world settings. These include: Theme 1: Power imbalances amongst stakeholders create substantial challenges and threaten ANI project success and ethical implementation. Theme 2: Insular decision making and stunted collaboration significantly jeopardizes systemic improvements to delivering ethical ANI-enabled health services.Theme 3: Current ANI-related ethics recommendations do not adequately capture the scale and impact of industry disruptions and what this means to the healthcare system and the patient. Theme 4: Commonly-used methods of ideation, communication, and implementation of ANI-related ethics policies limit the impact in real-world settings. Although there may be localized examples of ethical ANI outcomes on a project- / micro-level (for instance, through the use of algorithmic solutions), macro-level trends tend to be less favorable. Furthermore, the industry agnostic nature of these themes suggests that broader equitable ANI use starts far before a project begins and relates to the social, economic, and political contexts in which it is derived. Of which, a series of policy recommendations were developed for each theme through collaboration with subject-matter experts.