Should a norm arise among AI researchers that it is uncouth to talk about superintelligence or inquire into its possible risks, for fear of “giving ammunition” to critics, fear-mongers, crazies, or would-be regulators, then the recent gains of legitimation could quickly be reversed.
