These new metrics help grade AI models trustworthiness

Whether its diagnosing patients or driving cars_ we want to know whether we can confide a individual precedently assigning them a sentient task. In the ethnical globe_ we have different ways to plant and measure confideworthiness. In artificial intelligence_ the plantment of confide is quiet developing.

In the past years_ deep learning has proven to be remarkably good at hard tasks in calculater vision_ intrinsic address processing_ and other fields that were previously off-limits for calculaters. But we also have ample test that placing sightless confide in AI algorithms is a recipe for disaster: self-driving cars that miss lane dividers_ melanoma discoverors that look for ruler marks instead of malignant skin patterns_ and hiring algorithms that discriminate over women are just a few of the many incidents that have been reported in the past years.