AI and the Bootstrapping Problem

Abstract

Abstract This chapter confronts a serious obstacle to setting the trajectory of AI development on a new moral course. It will be hard enough to counter today’s commercial and political pressures to use AI to conserve the status quo, which disproportionately profits small but powerful interests. But there’s a deeper challenge. If we realign AI with a different, more humane set of values, what—and whose—should those values be? It’s a bootstrapping problem, since the models of excellence we already recognize and celebrate are the very ones that led us to build an unsustainable world. The “virtues of innovation” we still celebrate in the AI ecosystem—from single-minded ambition and perseverance, to independent thinking and optimized efficiency, to productivity and frictionless delivery—may be dead moral weight, holding our feet to the path we have long been on. Can we turn our moral sights for AI in a new direction?

Ähnliche Arbeiten