1 paper
Madeleine Pollack, Thomas Daillak, Mehmet Toner +3
We ask when an AI trained on historical decisions should be allowed to recommend decisions, defer to human experts, or in other cases, flag concerning patterns in data and deviate…