Artificial intelligence, invisible victims and the trolley problem

Abstract

The allocation of scarce healthcare resources inherently involves trade-offs between the interests of 'visible' and 'invisible' victims (ie, individuals who are aware that they are shortchanged by trade-offs and those who are not). At present, decisions regarding such trade-offs are often based on highly speculative predictions; the vast array of possible trade-offs simply cannot be enumerated, let alone the optimal outcomes calculated, by human beings. Artificial intelligence has the potential to change that reality by mining large data sets and other sources of information in order to produce far more precise and comprehensive predictions of likely outcomes and to delineate optimal allocation choices. Such technologies will inevitably render 'invisible' victims 'visible', generating a colossal, real-world trolley dilemma for anyone involved in medical or healthcare policy decision-making.

Ähnliche Arbeiten