Retiring the Human Race
Interview with robotics scientist Hans Moravec
by Dike Blair

Hans Moravec is a Principal Research Scientist of the Robotics Institute at Carnegie Mellon University. Born in 1948, he constructed robots even as a child (his light-seeking robotic turtle won him 1st prize at the Montreal High School Science Fair). In the 70s, while pursuing his Ph.D. at Stanford, he helped design planetary rovers for NASA, and was intimately involved with the information technologies exploding out of the university and industrial research labs on the West Coast.

In 1988, Harvard University Press published Mind Children. In it, Moravec outlines a history of robotics and artificial intelligence. By plotting the history of our computational power on a curve, and estimating the computational power of our biological brain, he makes the convincing extrapolation that our thinking machines will be our intellectual equals, and more, by the year 2040. His ordered and logical arguments build upon themselves in a manner that allows the reader to readily accept the plausibility of a future in which one can create a software copy of one's consciousness and download it, at the speed of light, into a mechanical "body" on another planet.

Hans was kind enough to let me preview his book in progress, The Age of Mind , which will be published by Bantam early in 1994 (see excerpt that follows this interview). Demonstrating a firm grasp of other disciplines--anthropology, biology, and socio-economics, he forecasts a wild near-future for human culture. The Age of Mind outlines what Moravec imagines will be the coming four generations of robots, their impact on society, and the eventual retirement of humanity as it moves into a post-biological state. Forecasting gets difficult when we consider intelligences greater than our own. As Science Fiction author Vernor Vinge describes the problem, "When this happens, human history will have reached a kind of singularity--a place where extrapolation breaks down and new models must be applied--and the world will pass beyond our understanding. A bright fellow like Mark Twain could predict television, but such extrapolation is forever beyond--say--a dog." Moravec envisions post-biological humans with super-minds whom he calls "Exes." They will live in space and inhabit strange bodies--like giant robotic bushes, resembling branching fractals, with trillions of atomic-sized fingers for the creation and manipulation of superdense particles. Moravec makes no claim to literal prophesy; however, the breathtaking beauty of his arguments and projections suggest that he is the Mark Twain of Vinge's analogy, and those who scoff at his ideas, the dog.

Dike Blair: In speaking of the near future, you predict problems, problems we are already experiencing, arising from unemployment as robots replace human workers. Could you describe your pension plan?

Hans Moravec: I see an incremental development. Even today there are not enough workers to support the increasing number of retired people. In the United States the writing is already on the wall--the Social Security System is not going to be able to support itself. The obvious answer to that problem is to subsidize it from general revenues. This provides a glorious opportunity because it means that taxes on companies can be used to subsidize retirement plans. If the Social Security system is modified to retire, in progressively descending age brackets, workers who have been excluded from the job market because of automation, we can gradually reduce the retirement age to birth. So, we will retire the entire human race and support it with industries which happen to be entirely run by robots.

Dike Blair: How do you see the Third World dealing with the radical elimination of the need for workers?

Hans Moravec: The thing that keeps the Third World out, right now, is the lack of appropriate skills in the population. The people don't have access to the information that would even give them the idea that they could improve their lives. As industry becomes more and more automated, the human factor becomes less and less of an issue. Even now, localities try to attract industry, even industry that doesn't employ many workers, because that industry will pay taxes. A local municipality may support itself by having a prison. I think the Third World will gradually have that option--some of it already does, like countries that have natural resources. Kuwait or the United Arab Emirates are good examples of this. Almost everybody is rich there.

Dike Blair: So you think that unbridled greed by the robotics industries is controllable?

Hans Moravec: There has to be a little bit of human collusion to make this work properly. I think international regulations could prevent the worst dangers for the countries that can't do it for themselves--something like the UN organizations, or the World Bank. If a robot company settles in a country, it would be required to pay taxes to the local human population. Something the technology can do rather easily is to provide an education to the local population, not that they would need it for jobs, but to bring them into the world community. Things like hand-held teaching machines could be part of the wealth distributed to the local population; and eventually, even personal robot tutors who would trigger all the instinctive learning inclinations that people have.

Dike Blair: What would you say to those who suggest that labor-saving technologies always promise a better life, yet it never pans out that way?

Hans Moravec: I disagree that it never pans out. Adjustments need to be made because of the problem of differential access to the benefits. There is a natural situation that has to be actively countered whenever you have a new technology, in the hands of a few, when that technology doesn't require a specially learned skill. In the early industrial revolution, the factory owners made out really well and the workers, terribly. The answer to that problem was found in the union movement which simply distributed the benefits of new technologies by distributing the work to more workers. I think that solution has run out of steam. What we need is what socialist societies have--redistribution. Despite my free-market reflexes, I think this is the only answer. There's no point in creating intense incentive for human beings to work harder and better when, even at their best, they are no match for these wonderful machines.

Dike Blair: This leads back to your pension plan.

Hans Moravec: I really see no alternative but retirement. The trick is to handle this retirement gracefully.

Dike Blair: Why is it that our artificial intelligence technologies seem more advanced than our robotics hardware?

Hans Moravec: I'm not sure that the hardware is really the problem. We can actually build pretty good legged and wheeled machines with pretty good sensors. The reason we don't have robots constantly moving around us isn't so much because of the mechanics, but because the computation isn't up to it.

Dike Blair: Would you describe how a mechanical process can reproduce our vivid mental life?

Hans Moravec: A robot that moves around the physical world encounters objects and thus needs to have an internal representation of those objects in a place we can call its mind. If it is interacting with other agents, be they other robots or human beings, it will need models with some predictive power of the behavior of those agents. This includes models of the psychology and emotional states of human beings with which it interacts. It's also natural for the robot itself to have an emotional life, although not every emotion a human being might experience is desirable in a robot. A sense of fear is almost certainly desirable so that if a robot encounters hints of danger in a situation it would act in a way which causes it to be more cautious, to move more slowly, to sense more thoroughly, or, perhaps, to move away from a situation altogether. A sense of love would be natural in a robot which is designed to serve a human being, because if it could estimate what makes its owner happy, and adjust its behavior to make its owner even happier. If one such robot were to see another such robot acting in this way, and applied its psychological estimates to it, it would say, "That robot is acting in a loving way." If the robot were to observe its own behavior in a third-person way, it would assign these emotional states to itself and could then analyze its behavior, to say things about itself like, "I was loving," or, "I was afraid." I think this rich mix of models of self and of the self interacting with the world gives you something very close to human consciousness.

Dike Blair: Let's talk about Science Fiction. It is apparent to me that you have influenced Science Fiction. Have SF writers influenced your work and who are your favorite authors?

Hans Moravec: I did most of my Science Fiction reading when I was young. I still read it, but it's a much smaller part of my life than it was back then. My mother was in the habit of accepting book club offers--you know, three free books and you can stop whenever you want. One of the offers was for The Science Fiction Book Club. This was my first encounter with serious Science Fiction--I was about ten years old. I got to pick three free books, but I didn't know any of the authors, so I picked the books on the basis of how thick they were in the illustration. I chose Childhood's End by Arthur Clarke, a trilogy by A.E. van Vogt that included The World of Null-A, and something called Astounding Anthology. I reread those books at least six or seven times (laughter). Later, I found more Clarke, and stumbled on to Issac Asimov and Robert Heinlein, and those all had a wonderful effect. The World of Null-A, however, had more of an effect than most--although the story isn't really so great (laughter). It includes a world in which social position is determined by something like an intellectual Olympic Games which are overseen by a giant computer, a thinking machine. This computer was the most real thing in this novel.

Dike Blair: It probably was the only character not in a loincloth.

Hans Moravec: I really identified with it (laughter). It was an extrapolation based on the first generation of electronic computers. It stayed with me--the notion of creating an intelligent machine really caught my fancy.

Dike Blair: Isn't the space elevator in Chapter 4 of The Age of Mind inspired by Arthur C. Clarke?

Hans Moravec: Well, in fact, I thought of it first (laughter). If you read Fountains of Paradise, there is a technical reference section in the back, and you'll see I'm cited there. I'd published a paper at the time Arthur Clarke was writing that novel which proposed a new version of the Sky Hook. I can't really claim credit for it, though, because it was a suggestion by John McCarthy (founder of the Stanford AI Lab) for which I worked out the mathematics.

Dike Blair: Would you describe your idea of a post-biological cyberspace?

Hans Moravec: The basic idea is that as the robots expand into space, the ones in the interior (around Earth) are surrounded by the space robots and they must compete, not through physical expansion but on the principle of their ideas being more interesting than their neighbors'. These interior robots are urbanized. They're like white-collar workers, rather than like the robots out on the edge, who are like frontiersmen. It's their ability to think that makes the difference. If they are made up of any inert lump of matter, that matter is better used for computation than for labor. Because of the competition, these interior robots gradually convert themselves into things that are almost pure computer--physical action goes by the wayside. Thus the whole interior space becomes a cyberspace. This continues to improve itself, and I have some very huge numbers--numbers that allow you to put 10 bits in the space of a human being. I'm trying to arrive at the ultimate limit to how much information physics allows you to store in a given space with a given mass. The numbers are huge--the cyberspace will be so smart that it won't need physical action at all, it will be able to do pretty much anything it wants. It will work with matter on a very fine scale and find very clever solutions for rearranging stuff.

Dike Blair: Are we beyond personality at this point?

Hans Moravec: Probably not, because personalities will be encoded inside the cyberspace. You could have these ethereal entities which think of themselves as relatively solid individuals who could link themselves with others pretty easily. It will be hard to draw strict boundaries around what is an individual and what is a community.

Dike Blair: Personalities will be able to choose their partners and form truly multiple personalities?

Hans Moravec: In a very fluid way that will change moment to moment. You may have a very liquid personality, or many personalities joining to form one greater personality.

Dike Blair: I think I should run the Voigt-Kampff Empathy test on you (Philip K. Dick's test for detecting Replicants in his novel, Do Androids Dream of Electric Sheep) . I'm going to outline a number of social situations. You are to express your reaction to each as quickly as possible. You're watching TV and suddenly you discover a wasp crawling on your wrist. . .

Hans Moravec: I remember that. When I was a child, I was very disappointed to realize that I wasn't a robot. It's a variation on the Ugly Duckling story--the child discovers that what makes him so different from other children is some wonderful latent power. I really wished that I was a robot.

The "Bekenstein bound" leaves room for a million bits in a hydrogen atom, in a virus, in a human being, for the Earth, in the solar system, for the galaxy, and in the visible universe. Chapter two estimated that a human brain equivalent could be encoded in less than bits. If it takes a thousand times more storage to encode a body and surrounding environment, a human with living space might consume bits, and a large city of a million human-scale inhabitants might be efficiently stored in bits, and the entire existing world population would fit in . Thus, in an ultimate cyberspace, the bits of a single human body could contain the efficiently-encoded biospheres of a thousand galaxies--or a quadrillion individuals each with a quadrillion times the capacity of a human mind.

Because it will be so much more capacious than the conventional space it displaces, the expanding bubble of cyberspace can easily recreate internally everything of interest it encounters, memorizing the old universe as it consumes it. Traveling as fast as any warning message, it will absorb astronomical oddities, geologic wonders, ancient Voyager spacecraft, early Exes in outbound starships and entire alien biospheres. Those entities may continue to live and grow as if nothing had happened, oblivious to their new status as simulations in the cyberspace--living memories in unimaginable powerful minds, more secure in their existence, and with more future than ever before, by being valued parts of such powerful patrons.

Moravec, Hans P., Mind Children: the future of robot and human intelligence, Harvard University Press.
The Age of Mind is scheduled for publication early 1994 by Bantam Books.