When to Trust Robots with Decisions and When Not To

robot decisions.jpg

A new piece in the Harvard Business Review suggests a framework for reflecting upon the implications of adaptive machine technologies such as autonomous vehicles. Specifically, a framework is suggested for evaluating which decisions people might be comfortable delegating to the algorithms.

While several of the subjective judgements made by the author in relation to the degree of “predictability” and to the “cost” of the cited examples appear highly debatable, the approach nevertheless provides a useful logical structure for organising debates about the design of our “friendly neighbourhood robots”…