After WWI, it was predicted that humanity would be extinguished by cropdusters blanketing cities with poison gas. After WWII, nuclear holocaust, for sure. After 9/11, terrorists with WMD, including suitcase and shipping container nukes. After the first nanotech, engulfment by "gray goo." The fact that these (and many other) end times prophesies failed does not mean that the new AI doom prophesy will fail - maybe this time it's different - but they should serve as a reminder of the cognitive bias of casually slipping from risk to existential risk, and to compensate accordingly, just as with other cognitive biases if one is a rationalist. John Mueller adds historical perspective here: Artificial Intelligence and the persistent survival of cosmic alarmism