The Ethical Consideration of Using Artificial Intelligence (AI) in Medicine

Method

Methodologically, we picked sensitive search words (finding the most crucial studies but with much irrelevant material) after the initial searches (more relevant research but with the risk of missing some important ones). During the article's analysis, we used the phrase “intelligent systems” in this section. The comprehensive, evidence-based examination covers intelligent medical systems. We chose recent publications since this is a rapidly expanding area. They have published it half in the previous five years. After evaluating abstracts, we estimated 35 texts.

We also searched on the Google search engine for 'Artificial Intelligence routine use.' This search returned another website (www.openclinical.org/aisinpractice.html) as being relevant. This site, which had not existed the previous year, held a similar list of artificial intelligence systems. An acknowledgment on this page said that Professor Coiera had initially compiled the archive; we used it with his permission; and that it was now maintained and updated by OpenClinical. The list was longer than the original website; OpenClinical had continued searching for systems. Professor Coiera's website listed the status of some 39 applications, while the OpenClinical website had increased this number to 50. Importantly, these numbers do not merely reflect the current systems in clinical practice. Professor Coiera's website also showed information on systems that were of “unknown status,” “under evaluation,” “decommissioned,” or “prototype.”

The OpenClinical website also listed all these systems except 'prototype'; the one 'prototype' entry from the previous list, ADE Monitor, is now listed as unknown. Of the 39 applications listed on Professor Coiera's website, we listed 25 as “routine use.” Only 12 of the 50 systems listed on the OpenClinical website were in “routine use.” Eight of the 12 systems listed as routinely used were exclusive to the OpenClinical list; we included only four of the systems listed as routinely used from Professor Coiera's list of 25. Two systems shown as being in routine use by Professor Coiera's earlier list were listed as being “decommissioned.” OpenClinical still listed the remaining 19 systems listed by Professor Coiera as being in regular service, but they had no status description. We assume that OpenClinical knew that these 19 systems were in routine use but can no longer be sure if they still are.

Of the 39 applications listed on Professor Coiera's website, we listed 25 as “routine use.” Only 12 of the 50 systems listed on the OpenClinical website were in “routine use.” Eight of the 12 systems listed as routinely used were exclusive to the OpenClinical list; we included only four of the systems listed as routinely used from Professor Coiera's list of 25. Two systems shown as being in routine use by Professor Coiera's earlier list were listed as being “decommissioned.”