To start with, AI won't demand human-like "sentience" to be an existential risk. Modern-day AI plans are given specific ambitions and use learning and intelligence to attain them. Philosopher Nick Bostrom argued that if 1 presents Practically any purpose to a sufficiently effective AI,