This article from the Bulletin of The Atomic Scientists site is an interesting overview of Nick Bostrom’s Superintelligence: Paths, Dangers, Strategies. The author rebuts Bostrom on several points, relying partly on the failure of AI research to date to produce any result approaching what most humans would regard as intelligence. The absence of recognizably intelligent artificial general intelligence is not, of course, a proof it can never exist. The author also takes issue with Bostrom’s (claimed) conflation of intelligence with inference abilities—an assumption the author says AI researchers found to be false.
AI artificial intelligence automation behavior bias bioengineering biology brain brain-computer interface brain functions brain hacking brain imaging cognition cognitive bias communication complexity consciousness data analytics economics embodied cognition emotions empathy ethics evolution evolutionary psychology existential risks fear gene editing learning machine learning memory metacognition metaphors neural networks neurons neuroplasticity persuasion political orientation psychology rationality reality singularity superintelligence synapses transhumanism