Contrary to the idea established in the collective imagination, the concept of artificial intelligence is not synonym with objectivity. An algorithm is not neutral: they are human creations, designed by teams of experts who specify the criteria for making a decision.
Based on this reality and as a consequence of the announcement of the Spanish Government considering algorithms to make decisions about patients with COVID-19, Dr. and Professor of Elisava Ariel Guersenzvaig signs an interesting article in the newspaper El País together with David Casacuberta (UAB).
In the text, its authors recommend a series of measures to be taken into account by the political and health authorities to guarantee fair and objective decisions in the field of health, a sector under enormous pressure these days as a result of the coronavirus crisis. In this context, what is important to avoid is, if necessary, a misuse of automated systems, such as algorithms for making decisions about patients in a hospital.
Among all the recommendations, it is important to highlight several points:
- To establish technical and ethical audits.
- To implement a final result of the algorithm that is not a simple score, but, for example, a series of criteria or informative dimensions that help health personnel make better decisions.
- To explain explicitly the criteria and rules used by the algorithm in order to understand and justify the reason for certain decisions.
Finally, both authors conclude that it is vitally important that the results of the algorithm are not binding: health professionals and not "the system" should decide.
Dr. Ariel Guersenzvaig is Director of the Master’s Degree in Design and Communication and Co-director of the Master's Degree in Web Design and Internet Projects Management, both programmes from Elisava.