Abstract
Background: Advances in artificial intelligence (AI) systems suggest that they can be used to improve healthcare outcomes via diagnosis, prognostication, patient management, risk assessment, etc. AI systems could be particularly useful in emergency healthcare (EHC) by synthesizing data to generate accurate conclusions rapidly. But the use of AI in EHC raises ethical, legal, and social concerns.
Objective: The present study undertakes a scoping review to collate, map, and synthesize existing literature on the ethical, legal and social issues (ELSIs) associated with AI in EHC. The aim was to assess which ELSI issues were recognized and analyzed in the current literature and which were under-explored.
Design/Methods: Online databases were used to identify papers published on the identified topic. An initial search strategy of IEEE, Pubmed, and Scopus yielded 156 unique records; 40 records underwent textual review, after which another 7 were excluded due to scope. The final 33 were analysed for content.
Results: Overall, the literature was mostly positive towards AI applications on EHC, with key themes aligning with the general AI ethics literature: transparency, bias, benefit/harm, justice, accountability, privacy and trust. Analyses of these issues, however, were mostly superficial and did not substantially engage with some of the distinctive features of EHC like urgency and high-stakes decision-making. In particular, urgency and stakes were under-recognised or under-explored in the EHC AI literature. Arguably, urgency in some emergency scenarios could justify more flexible ethical/regulatory standards, while conversely high-stakes contexts might require more stringent standards.
Conclusion: Lack of discussion of these contextual nuances suggests a significant gap in the literature of deeper research into the unique ethical, legal and social issues arising from AI use in EHC. This paper extends current knowledge by highlighting the need for deeper and more contextualized investigation of AI ethics in EHC.