Research Trends and Focus of Trust in Healthcare-Based AI: A Co-Occurrence Analysis and Modified Scoping Review

Abstract

BACKGROUND: As the potential of artificial intelligence (AI) in healthcare (HC) grows, so too does the number of potential risks, which has contributed to the development of regulations at national and international level that specifically address trust in and trustworthiness of AI. OBJECTIVE: Our research aims were 1) to assess the extent to which trust and trustworthiness are considered in the field of AI research in healthcare and 2) to show how the different scientific disciplines discuss both concepts. METHODS: We used a mixed-methods approach including a co-occurence analysis (19,940 publications) and a modified scoping review based on this, which included 30 publications. RESULTS: As shown in the co-occurrence network, trust and trustworthiness play a subordinate role in AI health-related research. 72 factors were identified that may have an influence on trust or trustworthiness. However, there is a lack of overlap in these factors across different disciplines, which is related to the heterogeneity of definitions and empirical approaches. CONCLUSION: The amount of research carried out to date regarding trust in and trustworthiness of AI systems in the HC sector appears to be significantly low and interdisciplinary approaches seem to be almost non-existent.

Ähnliche Arbeiten