Does Healthcare AI Meet Basic Ethics Principles?

By March 23, 2020

  • facebook
  • twitter
  • pinterest
  • linkedin

Ingrid Vasiliu-Feltes
Chief Quality and Innovation Officer
MEDNAX, Health Solutions Partner

Over the past decade we have noticed an exponential increase in the design, development and application of Augmented or Artificial Intelligence (AI) initiatives across many industries. Major thought leaders have been vocal about potential negative consequences or have promoted an even faster adoption due to the major beneficial impact it could have on our society. Healthcare AI poses even more nuanced challenges than other domains and several publications have highlighted the complex ethical questions that remain to be answered. The scientific community has been skeptical and some institutions even advocate the creation of certifications and /or credentialing for Healthcare AI.

Within the Healthcare AI spectrum there are numerous aspects that will need to be addressed, such as regulatory guidelines, intellectual property, privacy, patents, monetization to name just a few, however the most intriguing to me are the ethics of Healthcare AI.

So, does healthcare AI in its current deployed form meet even the most basic ethic principles?

Justice. One could state that the simple application of AI in healthcare would meet the impartiality and equality condition. Once we proceed a step further though with asking if access and ownership of AI data would meet this principle, it would already pose a major challenge.

Beneficence and Nonmaleficence. These two can certainly be addressed with appropriate governance and one could state that the most pertinent arguments would mirror those used for in the biotechnology domain. As a society, we continue to invest in R&D to find innovative solutions to many remaining healthcare problems, however those with maleficent intent can always aim to turn scientific progress into a biohazard or even bio-weapons and therefore effective counter-measures need to be continuously enhanced.

Confidentiality, Fidelity, Integrity. AI datasets, products or solutions will be extremely difficult to safeguard, however the same exact methods we currently use to protect all our healthcare data and need to continue to enhance should also apply to healthcare AI.

Autonomy. This one would definitely warrant a " depends" answer and it usually sparks controversial passionate debates with no conclusive answer.

As this cursory overview of the most basic ethics principles highlights, the answer is not as simple for Healthcare AI and we have a lot of work to do as a society. Despite all the challenges, AI is one of the strongest drivers for innovation in healthcare and we should perhaps all consider ourselves ambassadors or even promoters of Ethical AI in Healthcare.

Recent Posts