Thursday, March 31, 2011

The Global Brain Is Stupid

There is a common and oft-refrained motif in AI and Singularity related thought, the most recent incarnation being "The Global Brain (via H+). The Global Brain is generally defined as some sort of super gestalt entity consisting of the entirety of the internet, humanity, and/or both containing "distributed intelligence, one facilitated by communication and the meaningful interconnections between billions of humans, via technology."

Complexity does not automatically equal intelligence. Transmission of “information” does not necessarily equal intelligence. Human telegraph and phone networks transmitted great deals of information, and simply increasing the amount and even complexity of information transmission does not automatically make the system as a whole an intelligent gestalt with its own agency. Pure complexity is in fact selected against in evolution, as is illustrated in digital evolution simulations such as Polyworld, where the most complex-brained organisms wound up dying off. A brain's thought processes can be very complex, such as a paranoid schizophrenic's intractably intricate conspiracy theories of Illuminati-driven Area 51 cover-ups. We do not give these people Nobel Prizes and let them run nations precisely because they are spewing complex yet *unintelligent* gibberish.

Evolution does not “like” any given trait, traits only persist if they are selected for in a competitive environment, and it just so happens that intelligence worked out well in some — not all — environments in the history of Earth. The human brain evolved not simply complexity, but intelligence over millions or of years because of an arms race; that is, humans who failed to outsmart giant cats on the savanna died out, and so the most intelligent survived. I don’t see the “global brain” competing for survival against the ‘net gestalt of Jupiter, Saturn, the Kuiper Belt in a race to dodge black wholes or some such selective evolutionary environment. The global brain is not complex in order to make decisions to preserve itself, it is complex because it is beneficial for *humans* to transmit higher definition seasons of Lost on Netflix and sell penis enlargement pills to each other. If the global brain exists, its brain is literally made of 90% spam and junk mail. Can’t say I see much evidence of intelligence there.

"The wisdom of the crowd", Google's insta-answers, social media - these are not intelligent, despite the hype to pump the tech bubble 2.0 and appease nerd's sci-fi wish fulfillment or desire to attribute greater importance to their work than is really there. And the effects of these brave new technologies on our own intelligence are debatable. With a few keystrokes you can learn instantly the date of the invention of electricity, the barrel length of a Colt M4A1 Carbine, where the best tandoori chicken can be had in your area code, how to remove that pesky Disqus botnet trojan that's been using your computer to sock-puppet distort the tandoori chicken ratings and generally revise consensus reality under your nose. But how much smarter does instant question gratification actually make you? Basically, you're getting answers lightning-fast that you could've gotten with a bit more time if you went out and read actual books, talked to actual human beings, interacted with the actual world instead of just your touchscreen and Google's redirect to Wikipedia. A calculator can do math faster than any human mathematician, that does not make it *smarter* than even the dumbest human being. At the same time, if we over-believe the GB hype, we may become more like those quick-arithmetic'in gadgets; very fast at getting simple shallow answers, not so good at deeper complex analysis.

Author and polemicist Nicholas Carr points out that if you have a trivia hammer, you only look for trivia nails, and avoid deeper questions. His response to a recent study that showed college students answered questions three times faster with Google than using a campus library:

Lord knows it's great that we can answer well-defined questions a lot more quickly today than we could 20 years ago, and that that allows us to ask more, and more-trivial, questions in the course of a day than we could before, but Varian's desire to apply measures of productivity to the life of the mind also testifies to the narrowness of Google's view. It values the measurable over the nonmeasurable, and what it values most of all are those measurable variables that are increasing thanks to recent technological advances. In other words, it stacks history's deck. How did the University of Michigan researchers come up with the questions that they had their subjects find answers to? They "obtained a random sample of 2515 queries from a major search engine."

And let's not forget that Google, Twitter, these sites are not just "made of us", some benevolent, non-biased collective answer engine sourcing every human equally. These are corporations with bottom lines and shareholders who sell off reality to the highest bidder in the form of keyword auctions, "premium hit placement", and "promoted content" in order to make their money. The answers given to you by the "Global Brain" may not be the smartest answers but the ones which make the Network Owners the most money.

No comments:

Post a Comment