Great inventions, like beautiful women, are generally used first by cads.That is no argument at all against beautiful women, but it is half an argument against great inventions. With the inventions, folly persists, and bystanders get hurt. One has to ask whether the promises outweigh the threats.
Where robots are concerned, two fears stand out: first, that intelligent robots will put everybody out of work; second, that scientists are putting yet another gun in the hands of the baby.
You may recall that in Huxley's Brave New World, people were scientifically conditioned for the roles they would be assigned: the garbageman got his kicks working with garbage, the blast furnace workers enjoyed hundred-degree heat, people with boring jobs were too dull to be bored, etc. All this was arranged from the time they were embryos, so that they never became aware of any deprivation.
Hopefully, molding people to order remains as far out as it is offensive. It is evident, however, that if we can construct robots good enough to transplant human beings into, we can also construct a second class of robots that are just good enough to be willing slaves. With machines like that, what will be left for human beings to do?
If we ever did throw literally everybody out of work, we would at least gain some insight into the purposes of an economy: employment is not an objective. It's an input. Employment is sought -- and endured -- because it produces the goods and services which are the objective of an economy. To imagine that machines might take over all the work is to assume that they by themselves could produce all the things that people desire. It is to suppose all things superabundant and free, like sunshine. It is to imagine that we can invent an artificial Garden of Eden.
Fat chance! I for one prefer to let other people worry about "Whatever are we going to do with all our affluence, now that productive effort is only a conventional necessity?" In the world which I inhabit, affluence is not a problem; we are anxious instead whether we can continue to supply fresh water. If we did blunder into the Garden of Eden, there's soon be so many of us that we'd be eating the grass.
As long as human beings have wants over and above what the machines can keep up with, there will be work for men as well as for machines. This is not a matter for rejoicing. It's a permanent charter for wage-slavery. We haven't got it made, we're not going to get it made, and if we did we couldn't keep it that way. Even if we were not as good as we are at inventing new things to want, we would soon produce enough of us to make our more primitive wants pressing again.
Mind you, I'm not asserting that unemployment is impossible in the face of human wants. Monetary disorders can bring that about; so can political disorders, and maybe other things. Such disorders lay off machines as well as men. What I am saying is that "structural" unemployment -- the kind where business is booming, but people can't get jobs because they're no match for machines -- can't take in everybody.
The reason is obvious on the face of it: if something is being left undone, what's to keep somebody from doing it? And if machines can do it more cheaply than any available labor, why are not the machines turned to this neglected use, leaving their present area of employment to human labor? As long as there are unmet wants in the world, the machines must be too valuable in some employments to waste upon others. (By the same token, human labor must be too valuable in some employments to waste upon others -- but people never seem to look upon the matter that way.)
You may wonder, "But which work would be done by human beings, when a robot could be manufactured to do the task of any one of them?" The point is, just because it is technically possible to build a machine to do something, it doesn't automatically follow that it is economically attractive to do so. Machines, too, use up resources, both to produce and to run.
It's the old question of comparative advantage: a given machine may do the work of a hundred men, but if the same resources devoted to another use would produce ten percent again as much value, that machine is one too many. It will price itself out of the market.
There is nothing immutable in the terms of trade: if today a bulldozer seems more productive than its equivalent in laborers, tomorrow it may seem a gross extravagance. Why? Because fuels, metal, and fresh air are getting scarcer. As these scarcities make themselves felt in prices and taxes, the comparison will swing in the favor of laborers (with yokes of oxen, maybe). The same sort of considerations will determine the extent to which robots are used, and for what.
The real question is not whether human beings will lose their self-esteem and degenerate into vegetables as robots take over all the work. Relax about that, it's a temptation you haven't got. The real problem is just what it's always been: not that we will liberate all men from toil in a world of free goods, but that from time to time we will leave some unemployed in a world where toil is still needed. As long as it takes toil to produce things, it will also take toil to earn them -- and therein lies the potential for trouble. Total unemployment is an unattainable goal; partial unemployment is a genuine spectre.
It's a mistake to view the problem in the light of "man versus machine." Machines are owned by men. What you really see are different groups of men competing to service economic wants with different techniques. The mischief is done when some group hits upon a technique which other groups can't use. If the technique has impact along a broad front, it forecloses, for some people, all the alternatives at once.
The instance which is on everybody's mind when they discuss robots is, of course, the computer. Not everybody can be a computer programmer. (Neither can everybody be a brawny swordsman. Some of us scrawny, asthmatic intellectuals are well satisfied with the shift in advantages which has taken place since, say, William the Conqueror's time. Time was when we'd have had to renounce the world to stay part of it -- that was society's make-work for the physically deprived.) The boom market for computer programmers may or may not be putting other people out of work, in the aggregate, but it certainly alters relative statuses among people with different talents.
The main point to be noticed here is that such effects depend entirely upon human inequality. If people were truly equal -- equally strong, equally agile, equally quick on the uptake, equally whatever else might be needed --, then no productive technique could threaten anybody for long. It is because large masses of people feel they never will catch on and never could catch on that "automation" causes such widespread uneasiness. People apprehend that they are being placed at a permanent, irremediable disadvantage.
Under the circumstances, it's natural for people to feel that their only protection lies in being able to do something that machines can't do. This, as they see it, is all that stands between them and unemployment. The further extension of machine abilities puts them in an agony of insecurity.
Which brings us back to intelligent robots: I contend that the way out of our difficulties with automation is straight ahead. When intelligence is machine-replaceable as well as brawn and dexterity, nobody will have any special advantages. Artificial intelligence might even prove the greatest leveller of history, the greatest of all breaks for democracy.
To put the matter with less abandon: I have already argued that if machines could do anything a human might do, but could not in the aggregate meet all human wants, then there would still be employment for human beings. What I am saying now is that relative prices would determine to what extent machines replaced brawn, dexterity, analytic ability, and other human talents.
Machines take metal, plastics, rubber, silicon, etc. to produce; they take fuel to run, and create waste products to dispose of by air, water, or land. Different machines have different requirements; relative scarcities in these things, together with the details of technology, the comparative time-lags in production, etc., would determine the economic advantage of using each type of machine. Since all these circumstances could be expected to vary, so could the advantages accruing to different types of labor: those that go with each productive technique, and those that compete with each.
Actually, then, we could expect shifting advantages among different parts of the populace: now a little better for one group, now a little better for another. It's democratic in the sense that no advantages are systematic and permanent.
This is perhaps a surprising long-range outcome for an industry that likes to claim it is "freeing" people for "more human" uses. In the end, people will have pretty much the range of daily tribulations they have always had. And this is just as well, for the computer industry's sales slogan contains an unconscious arrogance: its definition of "human" is based on hopes, not on observation. The fact is that by this definition, everything "human" is foreign to the greater part of humanity.
Assuming, then, that intelligent robots won't throw everybody out of work, and if anything will tend to put everybody on a more equal footing, what about that other chilling aspect of the question: that we will be reintroducing slavery?
Everything I have said about robots and human employment assumes that the machines are property, and are used as such. If they have independent wants of their own, then they will indeed figure as competitors, and could replace human beings just as surely as Cro-Magnon Man replaced Neanderthal. Presumably, we will construct serving-robots without emotions -- or if emotions are necessary to intelligence, we'll give them the emotions appropriate to servitude. (I'm not talking here of the robots into which human beings get transplanted. These count as human beings, and human beings are their own progeny. The question is about robots used as machines.)
Doesn't this border on slavery? We know what was offensive about human slavery: the victims didn't like it. In Huxley's Brave New World vision of the future, the victims do like it, but we feel the situation is offensive just the same, because they ought to dislike it. These are human beings; it is immoral to tamper with them so that they fall short of what they could be.
Well, then, if we have the ability to make a robot fully human, but instead make it just intelligent enough to be a willing slave, isn't that immoral? Let your conscience be your guide. I myself find the dilemma uncomfortable; it is too much like the moral problem in abortion. I.e., when does a fetus cease to be a fetus and cross the line beyond which we regard it as a potential human? Up to what point are we justified in denying it its potential development? Up to fertilization of the egg? Up to three weeks? Six weeks? Six months? Retroactively after its first felony?
The problem will be the same with robots. If we build one that teaches kids and baby-sits, how can we justify not pushing it across the border into full humanity? But then, if we're honor-bound to do that, shouldn't we also upgrade the vacuum cleaner?
In the end, we'll have to settle for the arbitrary conventions which are the usual basis of morality (and the only basis possible). If we're wise, we'll forego making serving-machines with too close a resemblance to rounded human beings. Machines that are fairly specialized in function and appearance, even though some of the functions used to be tricks that only humans could do, will not bother our consciences excessively.
Finally, there is another way in which robots might make our lives miserable. Like everything else we make, they might be turned into weapons. "What a dream it was," sighed Orville Wright, as he watched the airmen of every country fly off to butcher other airmen. His lament speaks for all of us: in the twentieth century, everything we have done to realize the age-old dreams of Man has served in the end to make life more precarious.
Where robots are concerned, the probability that there will be military uses is something like 100%. It's useless to talk of control; if there are robots, there will be robot soldiers. The uses of science cannot be more noble than the purposes of society. The most control that scientists can talk about -- on their own -- is holding back inventions in the first place. Even that is not too practicable; the uncouth majority have ample means to make scientists do their will. (Like, you want to be a scientist, outside your spare time? Then you'll take the job that's offered.) Still, if the military uses of robots look particularly destructive, we might at least avoid a misplaced enthusiasm.
The fact that robots might kill does not of itself increase human insecurity. We have lots of ways to kill each other. Official assurances notwithstanding, we probably have the means to kill everybody. No doubt there will be some small increment in terror if we have machines that can move along the ground and make sure, but that is scarcely a major consideration.
If things ever get so far out of control that nations go all-out to annihilate each other, it won't be robots that seal our doom. Nuclear blasts, poison gases, and deadly plagues will make the big kills. (To be sure, the same technology that makes robots can make intelligent missiles -- but these are engaged in both delivering and intercepting warheads, so the balance of effects is unclear. Seeing that we are currently powerless to stop incoming missiles, and precise aim isn't necessary when populace is the target, smart missiles can't make our position much worse.)
The real military significance of robots is not as one more component in the Doomsday Machine, but as a highly usable "conventional" weapon in more "controlled" conflicts. It is said that a quarrelsome drunk is never too inebriated to recognize who is bigger than he is. Assuming that nations will be at least that sober -- for a while, at least --, and won't lob nuclear missiles at those who can dish it back, it is going to be important to us for some time what nations can do to each other short of mass slaughter. Under the cover of nuclear stalemate, we have already found it possible to carry on deadly-earnest wars, and the deadliness did not becloud anyone's perception that nuclear weapons were an unusable solution. (In the long run, over-confidence in this fact will be its undoing, but for now, it seems to work.) It may be that for a decade or two, Claymore mines and sharpened bamboo sticks will rack up a bigger score than the Bomb.
If we can count on continued rationality to hold off the high-powered stuff, then military robots will make a profound difference, indeed. The key to the matter may be seen in the phrase, "robot infantry." It should be obvious what a strategic difference it would make if intelligent robots could hold the ground.
By now we should know that the United States and its allies have one kind of military advantage, while the growing Communist camp has another. The advantages spring directly from the types of society. The relatively permissive, capitalistic society constantly gets the upper hand in weaponry, economic output, and innovation. The more regimented, socialized societies make it up in battlefield discipline, social cohesion, and continuity of purpose.
It's the same sort of mismatch that kept the Peloponnesian War going for thirty years. Democratic Athens ruled the sea, regimented Sparta was supreme by land. Athens tried everything once; Sparta was cautious but skillful. In the end, Athens did itself in; then, as now, public opinion was erratic, impressionable, and opportunistic. It was bold and brutal when picking on small islands, but panicky in the face of reverses. Ultimately a fatal mistake got made: Sicily was mistaken for a small island.
Everywhere the United States is supreme in the air; characteristically its adversaries rule the ground. The difference is no accident; it has to do with the relative advantages of productivity and discipline. Men can't fly, so the advantages in the air go to the more productive economy. Machines can't think, so the advantages on the ground go to the more disciplined soldiery.
Imagine, then, what a strategic difference it would make if intelligent robots could take the place of foot soldiers. Superior technology, faster innovation, and greater output would gain a decisive reward. Where this might lead in the long run is unclear, but the short-run implications are plain enough: it would be a windfall of power to the United States and its allies. To some extent, it would give a second wind to liberal democracy as the wave of the future. What confers military advantages on productivity must in passing favor free markets over command economies.
This argument should not be pressed too far. An advantage based on productivity has contradictory tendencies. It does smile on setups which give scope to the initiative of small groups and individuals. It does smile on economies which allow prices to register freely and quickly the relative advantages of things, and which allow buyers and sellers to respond naturally to the incentives thus created.
At the same time, an advantage based on production favors established power against revolution, rich power against poor, and stable society against unstable. This not what you would call good news for countries held under by other countries, for poor countries, or for countries with chronic internal divisions.
It is gospel in America to associate productivity with capitalistic incentives, individual freedom, etc. It only deserves to be a half-gospel -- say a New Testament. An older truth associates productivity with discipline -- you know, the Protestant Ethic and all that. The point is obscured for us because we make use of an indirect discipline: the market lays down socially acceptable guidelines to which the individual can conform in a manner of his own devising. It's not usable on the field of battle, but it gets us marching in good order for productive purposes.
Should anything break down this mechanism, we would soon find out that spontaneity offers no advantages in its absence. Private capital does have an Achilles' heel: it is extremely vulnerable to mob psychology. Whenever there is any serious question that the rules may change, that capital may be expropriated, that catastrophic events may wipe out wealth, capitalistic enterprise becomes paralyzed. The indirect discipline breaks down; individual initiative isn't worth a tinker's damn, for it is offered no opportunities to exert itself upon. And of course, the more the mechanism falters, the greater the chance that the institutions will indeed go down.
Because such episodes come and go in a capitalistic society, there is no guarantee that economic strength will always lie with the market economies. If the more authoritarian societies should manage not to be having a struggle for the succession at such a moment, or a "cultural revolution", they might well forge ahead in production.
History has other accidents. Robots are high technology. If "Middle America" insists on a vendetta against intellectuals, America may not even get the robots. Some of us remember the circumstances by which Germany was (fortunately) deprived of the A-bomb, while America gained it. American science won't thrive on anti-intellectualism, either.
So much for the military uses of robots. I am not fond of this aspect of life, but I see no way to prevent it. If robots must have military uses, I would just as soon the advantages accrue to freer rather than to less free societies.
Meanwhile, let's keep our eye on all sides of the bargain: the military uses of robots may have political implications (hard to predict, and which we may like or dislike), but they do not add appreciably to our existing capacity for destruction. The human species is already in mortal peril, and robots can't top that. Meanwhile, robots do offer a possibility of escape from that peril. They offer an indefinite extension of the human lifespan, and probably a means of dispersal to other planets. In the face of such possibilities, to weigh the robots by their strategic implications shows utter littleness of vision. Empires come and go; political systems are for an afternoon. Man's survival is another matter.
All the prices that are foreseeable are well worth paying.
|
|
Monsters, Slaves, and Honored Ancestors: The Three Lores of Robotics |
 |