Giving algorithms a sense of uncertainty could make them more ethical

This post was originally published on this site

Algorithms are best at pursuing a single mathematical objective—but humans often want multiple incompatible things.