When Will Computers Think for Themselves?

By Carl Weiss

Face it, since the 1980’s the personal computer has changed the world as we know it.  Before Apple and IBM started offering computers to the masses, the world was a much different place.  We weren’t as connected.  Life ran at a slower pace. The world seemed bigger.  But the advent of the microchip and everything that went along with it forever changed the ways in which we communicate, educate, shop, and do business.  In fact just about the only thing that hasn’t changed in the past twenty five years has been the fact that humans still control the destinies of every PC, tablet, smartphone, automobile, airplane and power plant.  Without programming, computers would be little more than diecast doorstops that are about as smart as an anvil. 

While science fiction novels and movies galore speak of the wonder and the horror of thinking machines, the fact is that there still aren’t any machines on the planet that can self-program or learn from their mistakes.  That’s the good news.  The bad news is that the day is coming sooner than you think when computers will be able to think for themselves. 

It all started with a game

Lëtzebuergesch: De Garri Kasparow géint de Computerprogramm Deep Junior am Januar 2003. (Photo credit: Wikipedia)

Computers are really good at games.  The reason is that games have rules.  Programming the rules into a computer is fairly straightforward.  Once programmed, a game playing computer has a distinct advantage over a human due to the fact that computers can perform hundreds of millions of calculations per second.  This was first brought to light in a big way when in 1997 the IBM computer Deep Blue beat the world’s chess champion Gary Kasparov. 

From Wikipedia - “Deep Blue, with its capability of evaluating 200 million positions per second, was 

the fastest computer that ever faced a world chess champion. Today, in computer chess research and matches of world class players against computers, the focus of play has often shifted to software chess programs, rather than using dedicated chess hardware. Modern chess programs like Rybka, Deep Fritz or Deep Junior are more efficient than the programs during Deep Blue's era. In a recent match, Deep Fritz vs. world chess champion Vladimir Kramnik in November 2006, the program ran on a personal computer containing two Intel Core 2 Duo CPUs.”

While an impressive feat, Deep Blue and its successors are extremely limited in what they can accomplish and how they can interact with humans.  All they do is play chess.  They not only are unable to hold a conversation about the nuances of the game, they don’t understand what the word nuance means.  However, all that changed in 2010 with the creation of the computer known as Watson.

IBM Watson (Jeopardy at Carnegie Mellon) - How I saved humanity! (Photo credit: Anirudh Koul)

Designed and built by IBM, Watson was designed to answer questions on the TV game show Jeopardy.  Unlike Deep Blue, Watson could not only understand the sometimes arcane questions posed on the show, but it could deliver its answers verbally. Relying on an extensive database of some 200 million pages of content contained in four terabytes of RAM, Watson competed on the air against two of the most successful human Jeopardy competitors of all time, Ken Jennings and Brad Rutter, beating them both for a prize worth one million dollars. 

“Since DeepBlue's victory over Garry Kasparov in chess in 1997, IBM had been on the hunt for a new challenge. In 2004, IBM Research manager Charles Lickel, over dinner with coworkers, noticed that the restaurant they were in had fallen silent. He soon discovered the cause of this evening hiatus: Ken Jennings, who was then in the middle of his successful 74-game run on Jeopardy!. Nearly the entire restaurant had 

Watson, Ken Jennings, and Brad Rutter in their Jeopardy! exhibition match. (Photo credit: Wikipedia)

piled toward the televisions, mid-meal, to watch the phenomenon. Intrigued by the quiz show as a possible challenge for IBM, Lickel passed the idea on, and in 2005, IBM Research executive Paul Horn backed Lickel up, pushing for someone in his department to take up the challenge of playing Jeopardy! with an IBM system.  Eventually David Ferrucci took him up on the offer.

In initial tests run during 2006, Watson was given 500 clues from past Jeopardy programs. While the best real-life competitors buzzed in half the time and responded correctly to as many as 95% of clues, Watson's first pass could get only about 15% correct. During 2007, the IBM team was given three to five years and a staff of 15 people to solve the problems. By 2008, the developers had advanced Watson such that it could compete with Jeopardy! champions.  By February 2010, Watson could beat human Jeopardy! contestants on a regular basis”

Can you say Doctor Watson?

More incredibly, after retiring from television, Watson was repurposed in 2013 to provide management decisions in lung cancer treatment at Memorial Sloan-Kettering Center.  IBM Watson’s business chief Manoj Saxena says that 90% of nurses in the field who use Watson now follow its guidance.  IBM is also looking at the possibility of using Watson for legal research,

While the software that upon Watson is based is available to large corporations and research centers

Deep Blue (Photo credit: James the photographer)

According to IBM, "The goal is to have computers start to interact in natural human terms across a range of applications and processes, understanding the questions that humans ask and providing answers that humans can understand and justify."  (such as Rensselaer Polytechnic Institute) The rub is that a system that meets the minimum requirements necessary to run a program as sophisticated as Watson currently costs more than one million dollars.  However, as computer chips become faster and less costly, it won’t be long before this kind of technology makes it to the masses.  When you realize that the computer power available in today’s smartphones is superior to that used to fly the Space Shuttle, then this claim is hardly beyond the realm of possibility.  (Even a handful of top executives at Google has espoused the goal of creating a computer much like the one on the series Star Trek.)

The Moore the Merrier

The real driving force behind the race to build intelligent computers did indeed start back in the 1960's with a tenet called Moore's Law.  Moore’s Law states that computer power doubles approximately every two years.  This little gem was coined by Intel co-founder Gordon Moore back in 1965, when he published a paper noting that the number of components in integrated circuits had doubled ever since their invention in 1958.  While this trend has slowed slightly over the intervening forty eight years, the nearly exponential growth of computing power has directly impacted every aspect of the electronics industry and brought us closer to the point where computers will be able to think for themselves.  It also led to other technological visionaries taking the next logical step.

“Peter Van DerMade, former IBM chief scientist, has spent over a decade studying the human brain and understanding how to replicate it in computer form. His new book, Higher Intelligence, tells the story of a 10-year breakthrough R&D project to build an 'artificial brain' chip that will help computers learn much like the human brain.  "By producing computer chips that allow computers to learn for themselves, we have unlocked the next generation of computers and artificial intelligence," Mr Van Der Made says.  “We are on the brink of a revolution now where the computers of tomorrow will be built to do more than we ever imagined.  Current computers are great tools for number crunching, statistical analysis, or surfing the Internet. But their usefulness is limited when it comes to being able to think for themselves and develop new skills," he says.“The synthetic brain chip of tomorrow can evolve through learning, rather than being programmed.”

Peter goes onto say in his book that he and his colleagues have already been able to simulate many of the functions of the human brain and convert them into hardware and software that enables computers to "learn new skills" without the intervention of programmers.  If he is correct, the next few years could see a paradigm shift that is more earth shattering than that of the advent of the silicon chip. Already we are seeing autonomous aerial vehicles flying the friendly skies and driverless automobiles plying the highways of Southern California.  With a few more iterations of Moore’s Law and a bit of tinkering, will we shortly be on the verge of intelligent systems, everyday robotics and machines that can outthink their makers?

If this isn’t quite the case yet, all I can say is, “Let’s come up with a game…”

Carl Weiss is president of W Squared Media Group, a cutting edge digital marketing agency in Jacksonville, Florida.  You can also interact live with Carl every Tuesday at 4 pm Eastern on his radio show, Working the Web to Win