the author of that article swerves to avoid the unpalatable truth: you can't create an artificial intelligence whose mind you can guarantee will be fenced in by passing social pietieshow are we going to teach AI using public data without incorporating the worst traits of humanity?
