Can medical algorithms be fair? Three ethical quandaries and one dilemma

刊登時間

The study by Kristine Bærøe et al. pointed out that three interrelated fairness quandaries and one fairness dilemma related to obtaining machine learning (ML) fairness in health are identified in this ethical analysis. Finally, there is the dilemma that arises from trade-offs between fairness and accountability.

  • the unfair data quandary
  • the unfair design quandary
  • the reasonable disagreement quandary

To avoid a rhetorical and insufficiently justified conception of fairness in ML technology, these fundamental and intangible challenges of fairness must be openly acknowledged and addressed. n addition, much more research on fair processes is called for to find ethically and politically sustainable responses to what fairness requires of ML algorithms employed in clinical care.

 

【MORE】