Saturday, October 14, 2017

How Hayek Understood That Artificial Intelligence Machines Will Never Take Over the World

Friedrich Hayek
By Andrew Sheng

There has been so much hype about AI that rock star historian turned futurist Yuval Noah Harari thinks that Man (Homo Sapiens) will evolve into Homo Deus – almost god-like through technology and science conquering famine, war and possibly even death (Vintage Books, 2015).

We used to suspend belief through science fiction, until large parts of the gadgets that we watched in Star Trek in the 1960s have all become reality.  With the ability of computers to beat the best human Go champion, science fiction is becoming reality.  The computer AlphaGo can not only learn from humans, but
play and learn from games against itself.  The machines are getting smarter than all of us.

The reason why Star Trek and other Sci-Fi movies are so popular is because they explore in an entertaining way all the moral dilemmas of science.  Will robots end up killing their makers?  When we intervene in a new planet, do we not kill off or change irreversibly their life forms?  As science and technology advances to the point where we can edit our genes, we can create either extended life or new monster Frankensteins.

The first thing to remember about AI is that it is a derivative of human intelligence.  AI is the result of cumulative human knowledge, with each discovery, each process, each innovation, each new institution leading to the opening of new fields of knowledge.  We are at the edge of an explosive field of new knowledge.

Current artificial intelligence is the discovery that knowledge advances through algorithms, defined as “a methodical set of steps that can be used to make calculations, resolve problems and reach decisions” (Hariri, 2015, pg. 83).  You can programme a computer or robot a set of algorithms, learning to learn and learning to adapt.  Perhaps computers can even learn how to identify and write algorithms.

The first thing you should remember is that human beings evolve through learning and adaptation.   That AI evolved from HI (human intelligence) is exactly like humans inventing the first dumb tool, then a smart tool.  Today, the smartphone is the “one device to rule them all”....paraphrasing Tolkien’s Lord of the Rings.  Indeed, the iPhone is already loaded with sensors to read maps, monitoring your heartbeat and steps for health and Siri can answer your dumb questions, with often dumb answers!   You don’t need a smart car full of things that you need to program from scratch.  You simply need to plug in your smartphone to rule your car.  Whether you master the full capacity of your iPhone is another matter.

The speed of AI and technology in general has been such that there are even those who think that this is end of manufacturing, and therefore the end of manufacturing in China.  Extrapolating in a linear fashion is the common mistake that futurologists make.   History does not evolve in a straight line, but is shaped by conflicting and cooperative forces the drive the evolution in a non-linear manner.

The present is a derivative of the past, just as the future will be a derivative of both the past and the present.  Events unfold in manners man has not been able to predict with accuracy.  Using Big Data and probability mathematics, AI enables us to extrapolate based on the best knowledge available, but there will always be an element of uncertainty - both known unknowns and unknown unknowns, including the unknowable.

In his 1945 paper, The Use of Knowledge in Society, Nobel Laureate philosopher/economist Friedrich Hayek understood better than any of his contemporaries that knowledge does not come coherently or rationally, but in “dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess.”  Thus,

Read the rest here.

No comments:

Post a Comment