04.11.2014 Views

Urgent Warnings, Breakthrough Solutions, Second Edition

Urgent Warnings, Breakthrough Solutions, Second Edition

Urgent Warnings, Breakthrough Solutions, Second Edition

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

configured their PCs with tens of billions<br />

of bits of RAM.<br />

If one accepts the comparison of<br />

computer bits to neurons as described<br />

by Moravec, then the computer<br />

’s growth in evolution expanded<br />

each decade what it took<br />

Mother Nature to achieve every<br />

hundred million years. Moravec calculates<br />

that human engineering of<br />

artificial intelligence is occurring at<br />

10 million times the speed of natural<br />

evolution.<br />

An approach to AI called embodiment,<br />

or embodied embedded cognition,<br />

maintains that intelligent behavior<br />

occurs out of the interplay<br />

among the brain, the body, and the<br />

world. Some philosophers, cognitive<br />

scientists, and AI researchers believe<br />

that the type of thinking done by the<br />

human brain is determined by certain<br />

aspects of the human body.<br />

Ideas, thoughts, concepts, and reasoning<br />

are shaped by our perceptual<br />

system—our ability to perceive,<br />

move, and interact with our world.<br />

Roboticists such as Moravec and<br />

Rodney Brooks (founder of iRobot<br />

Corp. and Heartland Robotics Inc.)<br />

maintain that, in order to achieve<br />

human-level intelligence, any AI-endowed<br />

system would have to deal<br />

with humanlike artifacts, and thus a<br />

humanoid would be the optimal robot<br />

to achieve this.<br />

The new field of evolutionary robotics,<br />

like its namesake of evoluseveral<br />

hundred years, it is really<br />

insignificant when compared<br />

to the glacial pace of natural<br />

evolution. In his 2000<br />

paper “Robots, Re-Evolving<br />

Mind,” Moravec compares the<br />

evolution of intelligence in the<br />

natural world with the progress<br />

occurring in the field of information<br />

technology.<br />

Natural intelligence evolution<br />

starts from wormlike animals<br />

with a few hundred neurons<br />

occurring more than 570<br />

million years ago. Very primitive<br />

fish that appeared 470 million<br />

years ago had about 100,000 neurons.<br />

One hundred million years<br />

later, amphibians with a few million<br />

neurons emerged from the swamps.<br />

One hundred fifty million years<br />

later, the first small mammals appeared<br />

and had brain capacities with<br />

several hundred million neurons.<br />

The bigger co-inhabitants at the<br />

time, the dinosaurs, had brains with<br />

several billion neurons.<br />

After the extinction of the dinosaurs<br />

65 million years ago, mammalian<br />

brains also reached sizes of several<br />

billion neurons. The first<br />

hominids of about 30 million years<br />

ago had brains of 20 billion neurons.<br />

You and I, and our contemporary<br />

human colleagues, have brains operating<br />

with approximately 100 billion<br />

neurons.<br />

Compare this to the artificial intelligence<br />

evolutionary track beginning<br />

with the first electromechanical computers<br />

built around 1940, which had<br />

a few hundred bits of telephone relay<br />

storage. By 1955, computers had<br />

acquired 100,000 bits of rotating<br />

magnetic memory. Ten years later,<br />

computers had millions of bits of<br />

magnetic core memory. By 1975,<br />

many computer core memories had<br />

exceeded 10 million bits, and by<br />

1985, 100 million bits. By 1995, larger<br />

computer systems had reached several<br />

billion bits. By the year 2000, a<br />

few personal computer owners had<br />

“[Hans] Moravec<br />

calculates that human<br />

engineering of artificial<br />

intelligence is<br />

occurring at<br />

10 million times the<br />

speed of natural<br />

evolution.”<br />

tionary biology, relies on the<br />

Darwinian principle of the reproduction<br />

of the fittest. This<br />

view posits that autonomous<br />

robots will develop and evolve<br />

from interaction with the environment.<br />

The fittest robots will<br />

reproduce by observing their<br />

interactions with the environment<br />

and incorporating mutations<br />

that increase their survivability.<br />

Humans will be unable to<br />

match the rapid evolutionary<br />

jumps afforded to completely<br />

artificial beings, even with advances<br />

in cybernetics and genetic engineering.<br />

Robotic humanoids will only be<br />

limited by the laws of physics and<br />

not by those of biology, which even<br />

genetic engineering can’t alter.<br />

Hopefully, the sort of destructive<br />

competition that eliminated the rivals<br />

to Homo sapiens in the past—including<br />

such competitors as Homo<br />

erectus and the Neanderthals—will<br />

not be repeated in the next evolutionary<br />

stage.<br />

In the best possible future, non-altered<br />

humans, humans with cybernetic<br />

implants, and robotic humanoids<br />

will learn from each other,<br />

borrow and share technology, and<br />

engage in friendly collaboration, cooperation,<br />

and competition to benefit<br />

all. In considering which robotic<br />

designs to support or, on the national<br />

level, to fund, that seems a<br />

good ideal to aim for.<br />

❑<br />

About the Author<br />

Steven M. Shaker is an executive<br />

in a market research<br />

and training firm. He is an<br />

authority on technology assessments,<br />

forecasting, and<br />

competitive intelligence. He<br />

is co-author, with Alan Wise,<br />

of War Without Men: Robots on the Future<br />

Battlefield (Pergamon-Brassey’s, 1988)<br />

and, with Mark Gembicki, of The WarRoom<br />

Guide to Competitive Intelligence (McGraw-<br />

Hill, 1998). E-mail steve.shaker@cox.net.<br />

<strong>Urgent</strong> <strong>Warnings</strong>, <strong>Breakthrough</strong> <strong>Solutions</strong> 21

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!