TOWARD THE RECOGNITION OF ARTIFICIAL HISTORY MAKERS

Abstract

ABSTRACTArtificial intelligence is a historical discipline. This does not simply mean that its history can be written. It is historical on account of its recursive basis for action: its systems turn to prior beliefs—often through multiple steps or layers—to make recommendations for the present or predictions for the future. Using the two rooms approach of Alan Turing's imitation game, I highlight the potential for machine and human histories to be recognized via at least the idea of weak artificial intelligence. This recognition illuminates the mixed nature of the logic of history, combining deductions and endoxa. Finally, I note that illuminating and exploring this mixed logic of history signals a turn to historiographical metaphysics with Aristotelian features and, thus, the recognition of histories by professional historians as only part of a historiographical world. This signals that the recognition of machine and human history makers does not simply turn on the acknowledgement of imitation.