| Summary: | Predictability is a prerequisite for effective human control of artificial intelligence (AI). For example, the inability to predict the malfunctioning of AI impedes timely human intervention. In this paper, we employ a computerized navigation task, namely, a game called lunar lander, to investigate empirically how AI's predictability compares to humans' predictability. We ask participants to guess whether the landings of a spaceship performed by AI and humans will succeed. We show that humans are worse at predicting AI performance than at predicting human performance in this environment. Significantly, participants underestimate the differences in the relative predictability of AI and, at times, overestimate their prediction skills. These results raise doubts about the human ability to exercise control of AI effectively — at least in certain contexts.
|
|---|