Pinker pushes sober AI safety over doomsday fears
harvard psychologist steven pinker argues doomsday ai scenarios are overblown, instead advocating for safety engineering. this challenges views like scott alexander's 20% risk assessment.
harvard psychologist steven pinker argues that fears of ai wiping out humanity are exaggerated, citing examples like the paperclip maximizer as unrealistic. he instead focuses on real dangers such as bioterrorism, cyberattacks, and unchecked ai agents. pinker advocates for sober safety engineering through independent oversight, liability, and human control. this stance contrasts with scott alexander's, who estimates a 20% chance of ai causing human extinction and calls for better safety research and international agreements to slow ai development. pinker's views are part of a broader debate on ai safety and risk assessment.