Neural networks are famously inscrutable — a computer can come up with a good answer, but not be able to explain what led to the conclusion. Been Kim is developing a "translator for humans" so that we can understand when artificial intelligence breaks down.