Tag
This paper investigates how uncertainty about AI prediction quality affects human decision makers' ability to benefit from complementary information, finding that negative error correlation between human and AI predictions enables robust improvement strategies.