When AI ethicists shoot from the hip: Stone-Age moral psychology in the age of AI

AI ethics has expanded rapidly, with a substantial share of its scholarly output arguably treating a wider-than-warranted range of AI-related developments as pressing moral problems. The paper argues that this can, to some extent, be explained by three evolved cognitive mechanisms: the overdetection of agency, the overmoralization of affectively charged situations, and the overpropagation of morally salient information. It further suggests that these mechanisms are likely at work among AI ethicists under conditions particularly conducive to their activation, including the genuine novelty of many AI technologies and the moral questions they raise, the limited protective effect of ethical expertise against certain cognitive and moral biases, and a professional environment that rewards the identification of ever-new moral problems. The paper develops a conceptual framework comprising a set of arguments and explanatory hypotheses aimed at explaining why some AI-related concerns proliferate and intensify in ways disproportionate to the actual risks and capabilities of existing AI systems.