’Bout My Computer Generation
increasing in sophistication at a pace that outstrips biological
evolution, and scientists working toward creating the self-aware
robot, are humans in danger of becoming last year’s model?
By William Kanapaux
All the hype about the philosophical underpinnings of The
Matrix Reloaded overlooks the movie’s obvious shortcomings
when it comes to depicting artificial intelligence: First,
humans would make a lousy energy source compared to plutonium,
and, secondly, if an artificial intelligence were smart enough
to create a virtual world for humans kept alive in suspended
animation, it should be able to write programs that shoot
Chances are, the real-world future of artificial intelligence
won’t be anything like the versions depicted in The Matrix,
the Terminator films, 2001: A Space Odyssey or
A.I.: Artificial Intelligence. Depending on the expert,
artificial intelligence could be a benign presence humming
along in the background of our lives, a potentially destructive
force that we can only hope doesn’t fall into the wrong hands
or a new life-form that quickly evolves beyond anything we
could ever conceive.
Artificial intelligence is already with us, in modest form.
Speech- and face-recognition software and programs that monitor
credit card transactions for fraud are examples of artificial
intelligence that we’ve come to accept as a part of life in
the modern world. One such project is Cyc (pronounced “psyche”),
a $60-million effort to teach computers common sense through
software that would become standard equipment on all machines.
The Cyc team has been at work since 1984 developing an “inference
engine” that would serve as the brains for machine intelligence.
Its software so far has more than 1 million rules that are
designed to help machines understand those things about the
world that we take for granted: Once people are dead, they
stop buying things. Glasses of liquid should be carried right
Its creators also instructed Cyc to ask questions if it needed
a concept clarified. In 1986 it asked whether it was human,
and later that year whether other computers were involved
in similar projects.
The Defense Department found Cyc interesting enough to invest
$25 million in it for testing as an anti-terrorism tool. Meanwhile,
the Defense Advanced Research Projects Agency (DARPA) is launching
its own effort to create a cognitive system that possesses
Most computer systems, while powerful, are complex and susceptible
to crashes. DARPA’s goal is to create a system that can take
care of itself—testing for software problems and containing
them before they crash the entire system. It also would be
able to learn from experience, identify relevant information
within huge amounts of data and make plans by anticipating
different scenarios. While its battlefield implications are
obvious, such a system also could be used for cleaning up
toxic spills and aiding in disasters.
But overall, the AI field seems to be moving at a creaky pace,
with no HALs or Agent Smiths in sight. Instead, we have a
small robot vacuum called Roomba that can clean rooms unattended—at
a snail’s pace—and sells for $200 at the Sharper Image. And
the MIT Artificial Intelligence Lab is busy with Cog, a humanlike
robot with two arms that researchers hope will learn to sense
its own physical actions and determine causal effect.
While the pace so far seems glacial, some top experts are
concerned that progress may soon exceed our ability to comprehend
or control it. Research in areas such as computer vision and
navigation will eventually merge with discoveries in machine
learning and reasoning, leading up to the creation of fully
autonomous thinking machines. The ongoing exponential growth
in computing power will allow these machines ultimately to
develop “brains” beyond anything a human can imagine.
Except for the part where things turn out badly.
In 2000, Bill Joy, cofounder of Sun Microsystems and a creator
of the Unix operating system, caused quite a stir with his
dire predictions in “Why the Future Doesn’t Need Us.” Published
in Wired magazine, it offered a chilling version of
where the future of technology might lead. In spirit, the
Matrix films aren’t too far off the mark.
The essay predicts a future in which self-replicating intelligent
robots—some operating on the molecular level—will crowd out
humans and forever alter life as we know it. Humans will either
merge consciousness with the new machines or face extinction.
Sure, it sounds like the stuff of science fiction, but to
Joy and other futurists, these events are only 30 to 50 years
away. “By 2030,” he writes, “we are likely to be able to build
machines, in quantity, a million times as powerful as the
personal computers of today.”
Once intelligent robots exist, Joy says, it’s only a small
step to a robot species. Robotics researcher Hans Moravec
says that on an evolutionary scale, computers currently are
no smarter than the fish that first crawled out of the primordial
sea. But they are evolving 10 million times faster than biological
creatures and should overtake humans in 50 years. As for the
outcome of this close encounter between humans and their superintelligent
progeny, he observes, “Biological species almost never survive
encounters with superior competitors.”
Moravec, in his book Robot: Mere Machine to Transcendent
Mind, envisions the evolution of extraordinary forms of
machine life that expand into all corners of the universe.
He paints a wild vision of robotic races far removed from
human experience, evolving in all directions. Inventor Ray
Kurzweil, author of The Age of Spiritual Machines,
sees a more benevolent future. He believes that humans will
achieve immortality by merging with robots, and that we should
welcome these intelligent machines as our offspring.
In at least one respect, our future will resemble The Matrix,
Kurzweil predicts: By 2030, nanobots will course through
our bloodstreams, mapping brain functions and sitting at our
nerve endings ready to create any virtual reality we wish.
Humans will be able to activate and deactivate these nanobots
at will, moving freely between the real world and the virtual
Clearly not everyone subscribes to these visions. It would
take more than computing power to create an intelligent, autonomous
organism, the thinking goes, especially one possessing some
form of consciousness. And it would seem that humans, as the
creators of this intelligence, would have some kind of control
No doubt it all seems far-fetched. But think about the way
technology has crept into our lives. Fifty years ago, a mainframe
computer as powerful as today’s desktops would fill a warehouse.
Today, they’ve “invaded” our homes by the millions. Or think
of The Matrix Reloaded. A nonstop assault of special
effects creates an illusion of reality compelling enough to
provoke a physiological response. Adrenaline pumps through
our veins. Our muscles tense. Our bodies feel the action,
even as we sit immobilized in a darkened room.
Maybe we’re closer to these outlandish futures than we think.