Artificial Intelligence And Clever Hans
Why we don’t need to fear an AI Apocalypse
We humans evolved under relatively simple conditions in which abstract reasoning was of little value. That’s why even today most people struggle when attempting to reason. Syllogisms are an easy way to demonstrate the limitations of human thought (though there are a great many equally embarrassing ways), and so when we speak of “intelligence” we need to remember that we don’t actually possess very much of it. What little we do possess, we rarely use because conserving calories was Job One for 99% of our evolutionary history and attempting to reason burns precious calories. Thus, we avoid it whenever we can. That’s why most people watch entertainments on their flickering screens rather than studying advanced mathematics, or physics, or learning a new language.
What was extremely important to us as a primate group species was to infer feelings and intentions in others. If our ancestors could avoid conflict with other group members by having an adequate theory of mind (the ability to imagine what’s happening inside the brain of another person), or make alliances by inferring shared interests, this could confer significant advantages. So over time humans became better and better at inferring internal states and intentions — so much so that we automatically project it onto nearly…
