Would you agree with an artificial intelligence algorithm deciding which patients should receive donated organs in transplants?

For @Alvy — August 7, 2022

Photo (CC) Camilo Jimenez @ Unsplash

I saw a reference to an article written by researchers from the Department of Artificial Intelligence Applied to Medicine at the University of Cambridge who published an article explaining how they found a method for select patients who should receive donated organs using a artificial intelligence algorithm this improves on the current human criterion, which, by the way, is also based on various mathematical models, of course.

Conclusion: The researchers found that the machine learning algorithm (called Organs) was able to produce better results in the simulations, which included a mixture of real data from the past 26 years and carefully balanced imaginary data. these things are complicated for ethical reasons and because many factors must be taken into account: the type of organ, its rarity, the years of life it can offer the recipient, the possibility of rejection, etc. The new algorithm manages to make a somewhat peculiar selection that maximizes the total number of years of life of all the beneficiaries, something that sounds like a good idea and something socially acceptable without a problem.

What they observed was that the algorithm learned to offer the new available organs not only based on who can benefit the most, but also based on the probability that a new similar organ will appear in a short period of time (a factor called “organ shortage”). This benefits those who need an organ with certain “rare” characteristics when you know that other compatible organs that don’t have it (and don’t need it) are likely to appear for more. other patients who could have priority over the first.

The work is from a few years ago, but still they talk about him. In reality, Organs is not the only one: there are others like OrganSync, OrganBoard there iTransplantation that focus on the same area of ​​medicine.

It seems to me that if a mathematical model is currently in use and this type of algorithm is an improvement or an extension of the previous one, why not leave the decisions in their hands. In the end, it can be justified that the result is better for society as a whole. Of course, this also shows the importance of explainable artificial intelligences (XAI) able to properly account for the motivation of their decisions.

{Photo: (CC) Camilo Jimenez @ Unsplash.}

Leave a Comment