Dylan Evans critiques the preoccupation with the dangers of unfriendly superintelligent AI, suggesting that it is a concern more likely to afflict highly intelligent individuals. He argues that while superintelligent machines do pose potential risks to humanity, there are more immediate and probable threats that should command our attention.
Evans introduces the concept of “Pascal’s mugging,” where individuals are convinced to take unlikely bets with large potential rewards by framing the outcome in terms of vast future utility. He suggests that this line of thinking can lead to diverting resources away from more pressing issues.
Evans posits that some people may be drawn to the idea of being the saviors of humankind, a notion that appeals to their narcissism. He also suggests that there are economic interests at play, with individuals and organizations benefiting financially from promoting the idea of unfriendly AI as an existential threat.
In essence, Evans warns against being overly fixated on the risks of superintelligent AI to the detriment of addressing more immediate and tangible challenges.