One of the most classic ethical dilemmas of medical practice is that of the organ transplant, typically a liver or kidney to go to the lesser of two abusers of their own bodies when discussed at the dinner table. The real plight that many people have to face is the fact that organs available for transplant are few and far between, so committees have to decide who gets first rights to various organs. Now, an already complex medical dilemma is becoming even more complicated with the prospect of artificial intelligence possibly determining who gets any given organ.
Researchers at the University of Maryland and Duke have engineered a new algorithm that aims to more than adequately account for the ethical decisions human beings ordinarily make based on their responses to surveys regarding who deserves a hypothetical kidney. “We’re not advocating that AI replace human decision making,” Jana Schaich Borg clarifies as a Duke social science professor who helped engineer the algorithm. “AI can inform and improve decision making. It helps unearth some of our biases, brings them out onto the table so that we can discuss them and be more reliable.”
Making this algorithm started with crowdsourcing participants for a survey via Mechanical Turk, Amazon’s online crowdsourcing marketplace. The research team did this to figure out which aspects are most relevant and acceptable for determining which patients get first rights for organ donation. The survey helped them quickly rule out several attributes quickly like race, which most respondents agreed was an inappropriate determinant. On the other end of the spectrum, whether or not a patient classifies as a dependent proved disagreeable enough that the team knew to table that for a later phase of study.
The study dealt specifically with a prospective kidney transplant, so the three determining factors that the team found to be most contentious due to all manner of lifestyle baggage as well as presumed, remaining life expectancy were drinking habits, age and cancer history. The team proceeded to engage the second set of some 300 Mechanical Turk participants for another phase, and the study rapidly accrued copious amounts of data. Ultimately, they produced an algorithm that decides between any two candidates which one should receive the kidney by weighing several of the tested questions from survey phases.
The resultant algorithm that most Turk respondents ended up favoring, the algorithm favors young people who don’t drink or only drink occasionally and who have no cancer history. Obviously, this is a common rationale, but that an algorithm comes out that way based on survey responses from people progressive and tech-savvy enough to use Mechanical Turk in no way negates the fact that this same algorithm waitlists patients who are older and sicker — the impetus for the common rebuttal.
Walter Sinnott-Armstrong, a Duke philosopher of ethics who contributed to the project, says that those whose job it is to make these decisions sit in committee seats are already doing what the algorithm is doing and arguably inspire no more confidence than it does. He says the task of making these decisions in the first place is inherently a job in which ethical decisions are made by people who can’t adequately explain their choices any better than the algorithm does, and the algorithm represents hundreds of survey respondents who similarly can’t explain. “With enough subjects, you can get much more information about how they’re weighing those things,” Sinnott-Armstrong explained.
“You can find out, for example, are they taking the fact that maybe it’s this person’s fault that they are in this situation because they drank so much [into consideration]? How does that weigh against the fact that they’re older, which is not their fault, or that they had skin cancer, which is not their fault?” Whether or not fault is a viable factor or should be is a point of contention in and of itself. A 2016 study greatly catalyzed the progress of an alternative means of dealing with these bioethical dilemmas, though. For example, it’s credited with inspiring the development of a method for cultivating bioartificial kidneys.
This sparked a slew of studies and research projects aimed at generating synthetic organs as a way to bypass the organ transplant dilemma altogether. The most controversial and earliest of these methods is that of stem cell research. Stem cell applications remain heavily restricted for a myriad of ethical reasons, but studies have already proven that stem cells can more than adequately grow organs in vitro for implanting. Beyond stem cells, though, many scientists have turned to the wonder of 3D printing to produce artificial organs now as well. In these contexts, it’s often called bioprinting, and researchers have already bioprinted thyroid glands, beating heart cells and a tibia replacement that’s actually implanted into a patient’s body already.
Organ transplant research, however, is stifled in many regards by all manner of regulations and laws that prevent further understanding. On top of that, once a patient’s brain is dead, physicians have very little time to conduct all the research they want on the body and its organs because brain death no longer coordinates tasks that keep a person alive; ergo, studying the lungs or the heart begins to rapidly lose context when nothing else is working. “You can see the kidneys don’t work as well, the lungs don’t work as well, and the heart can look stunned,” according to trauma surgeon Darren Malinoski at Oregon Health and Science University.