Personal but Necessarily Predictive? Developing a Bioethics Research Agenda for AI-Enabled Decision-Making Tools

Abstract

Any human or AI system ambitious enough to mine my every social media post, purchase, or Internet search history is likely to infer some very accurate details about me.I am in my 30s, an avid traveler and consume an unhealthy amount of espresso daily.However, no medical record entry, prior purchase history or social media post used to train a P4 meant for me would reflect the reality that I was forced to make substituted judgements for immediate family members teetering on the edge of death.Twice.These life experiences, the decision-making hardship in the moment, and the enduring memories from those events, while undocumented and unseen, are perhaps the most accurate datapoints that should be used to encode my preferences and decision-making processes if AI were ever to make substituted decisions for me.

Ähnliche Arbeiten