Intelligence is not Artificial

Why the Singularity is not Coming any Time Soon And Other Meditations on the Post-Human Condition and the Future of Intelligence

by piero scaruffi
Cognitive Science and Artificial Intelligence | Bibliography and book reviews | My book on consciousness | Contact/feedback/email

(These are excerpts from my book "Intelligence is not Artificial")

Dystopianl Quartet 1. Andante: Lessons Learned from A.I.

The A.I. scientists of the 1950s have been mocked for making unreasonable predictions about the intelligence of computers. It turns out that they were mostly right. Those scientists lacked machines powerful enough to implement the theory, but, decades later, the theory can be implemented on powerful GPUs (especially when munificent benefactors like Google volunteer to buy thousands of them) and it works. Neural networks made of these powerful GPUs can indeed simulate a lot of "intelligent" tasks. Therefore A.I. scientists have proven to neuroscientists that a lot of our "intelligence" is simply computational math, algorithms, formulas, very intricate systems of equations. However, some tasks are too intelligent for us and still too intelligent also for the machine. For example, the original app: weather forecast. We still can't accurately predict the weather (except in sunny Silicon Valley). Weather forecast involves too many factors, and we don't have the dataset of past behavior that could help "recognize" the pattern for the future. There is a second limitation to machine intelligence, and sometimes to human intelligence: lack of common sense. Common sense requires an almost infinite amount of knowledge about ordinary things. It turns out that the ordinary is much more difficult to map than the extraordinary. It is not difficult to describe mathematically how a master plays chess or weichi/go. It is very difficult to describe mathematically how a waiter cleans a table at the restaurant. Common sense is hard to replicate with algorithms. It has proven much easier to remove the need for common sense. To a large extent, the history of human civilization is the history of removing the need for common sense. We don't need to run away from tigers anymore and we don't even need common sense to cross a very busy street. We have structured the environment so that life can be easy, safe, and predictable. In the "developed" world, there are rules and regulations for just about everything. We train children from a very early age to abandon common sense and follow rules. We strive to remove every possible glitch from our factories, subways, malls, and streets. Removing the need for common sense enables even the dumbest people (like me) to easily survive. I can write, teach, travel and enjoy my hobbies regardless of whether today my mind is sharp or not: there is a vast system of rules and regulations that takes care of me. We turned human behavior and the human environment into a well-organized, highly predictable, dumb, machine-friendly system. No wonder that machines can easily coexist with humans. The machines don't have to be very intelligent to deal with people because people have built an environment that doesn't require the one thing that machines don't have: common sense. And no wonder that we can easily replace humans with machines: humans have been trained all their life to behave like machines. So the issue now, as far as business is concerned, is not whether a machine can do the same job (it can, given a well-structured environment) but which one is overall cheaper to "operate". We transferred “intelligence” from the human brain to the environment so that humans don’t need intelligence anymore to survive in the environment; but that also means that machines can now do what only humans were capable of doing when the environment was chaotic and surviving in it required intelligence.

Back to the Table of Contents


Purchase "Intelligence is not Artificial"
Back to Cognitive Science | My book on consciousness | My reviews of books | Contact