instead of implementing humanity’s coherent extrapolated volition, one could try to build an AI with the goal of doing what is morally right, relying on the AI’s superior cognitive capacities to figure out just which actions fit that description. We can call this proposal “moral rightness” (MR).
