AGI vs Human Intelligence
On the constraints of human intelligence
When describing what artificial general intelligence (AGI) would look like, people tend to compare it to the way humans think. Implicit in that comparison is the idea that human intelligence is broad and general, that is, fit to tackle a broad range of problems and situations. This is, of course, mostly true: human cognitive abilities are not narrow and limited like today’s computer algorithms (current artificial intelligence).
However, there are parallels between our cognitive abilities and how we perceive the world through our senses. Evolution has shaped the way we perceive reality in order to optimize the Darwinian fitness of our species. This allows us to perceive certain things, while many others remain undetectable.¹
Our cognitive abilities and ways of reasoning have been similarly shaped by evolution to increase our chances of survival and reproduction, and not necessarily to be always logical or coherent. It’s not just that we need to make a deliberate and strenuous effort to master certain concepts that do not come naturally to us, e.g., calculating probabilities. There are certain things, particularly those related to us, such as the meaning of our existence, our role in the universe, our essence, our morals and our actions, for which evolution seems to have programmed in us nearly inescapable biases and even cognitive blinkers. Such programming effectively blocks us from reasoning rationally, and from seeing certain things. The resulting irrationality has little to do with what we normally call intelligence, or with what IQ tests measure. Extremely smart people can still be wholly irrational and entirely misguided with respect to certain topics, especially those connected to humankind and to us as individuals, as said above.
One has to wonder how many “unknown unknowns” there are in our cognitive blind spots: things we don’t know that we don’t know, or worse, “things we know for sure that just ain’t so” — notions and ideas that evolution has “deliberately” placed beyond our cognitive grasp. Let’s hope that if we ever develop AGI, it will not be similarly blinkered, and then perhaps some of those “unknown unknowns” will be revealed to us. On second thought, perhaps it would be better for our happiness and prosperity if nothing of the sort were ever revealed…
¹ Some go so far as to argue that even what we do perceive does not reflect reality at all. That seems quite a stretch. Even though what we perceive may often be distorted (and occasionally outright wrong) or filtered through our brain so that certain things are emphasized or ignored, it does for the most part reflect the real world. It’s not just an endless hallucination that has no bearing on reality. Our perceptions may be narrow in the sense that they detect only those aspects of reality that are relevant for our survival and reproduction, but there is no compelling reason to surmise that they do not represent the real world to a great extent.
