Wednesday, January 12, 2005

Transhumanism and economics, with a long note on Artificial Intelligence

Elaborating on the thread mentioned in the earlier blog (short intellectual auto), I would ramble a bit on the relationship between the transhumanist technologies (artificial intelligence, cybernetics, genetic engineering, nanotechnology) and economics.

First a brief historical survey:

Adam Smith, the father of modern economics, lived at the dawn of the industrial era (the late 18th century), and many of his most salient observations--notably of course about the division of labor--can be traced directly to the practices of the early industrialists. Adam Smith's economics is the fountainhead of the 'classical' stream of economics. Based on their own economic milieus, Ricardo and Malthus will elaborate Smith's theory in different forms, while JS Mill will give it a supreme elevation somewhere in the middle of the 19th century.

In the 20th century, Alfred Marshall will incorporate the insights of the 'marginalists' (e.g. Jevons who came up with the theory of marginal utility) and earlier economists in a grand neo-classical synthesis. His student, JM Keynes will use this foundation to build up his General Theory in response to the greatest crisis in capitalism: the Great Depression, while in the later part of the 20th century, neo-classical economists like Milton Friedman will also derive crucial sustenance from Marshall's ideas to react to their inflationary era.

The economics of Adam Smith and his later disciples is fundamentally one based on the more or less free operation of the market (how free depends on the economist in question)--a market made so much larger and more all-encapsulating by the industrial revolution and its implications. It is also allied basically to a political view that accepts individual liberty, at least in questions of private property and market choice. It is thus usually congenial to democratic or social-democratic systems. The Fascist and Nazi experiments have not operated long enough (thankfully) for us to observe how well a market economy can function under such capitalist totalitarian regimes.

But there is of course a totally divergent form of economics: Marxist economics. Karl Marx worked in the 19th century, during the height of the industrial revolution, and his insights, like that of the classical economists is linked to the economic phenomena of the day. By contrast to the classical economists, Marx (among many other differences) gave much more emphasis to labor as the main factor of production, downgrading the status of capital and land. His downgrade of capital and capitalist practices naturally led him to favor a 'command economy', where the market mechanism is replaced more or less by State direction. This form of economics is supposed to culminate in a form of utopian anarchism, though the actual historical form it took was communist totalitarianism.

From this brief historical survey, one can see how economists usually work out their systems in response to the prevailing economic problems of the day. Different economic conditions and tensions in turn require, inspire and reinforce new systems of economics and politics.

And of course a key factor driving economic change is technological advancement. For instance, the massive productivity growth of the industrial revolution is made possible by technological innovations like the steam engine, railways, telecommunications, electricity and mass production. Hence we need not be surprised if our present day technological advancement leads us towards a revolution in our economy, and thus in our economics and politics.

Transhumanism and economic change:

We, like Marx and Adam Smith, lives in a revolutionary era--albeit in its earliest stages. Perhaps future historians will term the 21st and 22nd centuries not as the nuclear age, space age, genetic age or computer age, but simply the transhumanist or 'post-human' era. Industrial era technologies empower a vast increase in labor productivity through mass communications, transport, production and so on. It is dependent on a symbiotic relationship with machines--a relationship closer and more dependent than any hitherto seen.

Transhumanist technologies like genetic engineering, artificial intelligence and cybernetics, can accomplish this as well, but also exponentially more. Artificial intelligence for instance may eventually carry the industrial logic of higher production for less laborers to its finale: the production of work for zero laborers (I do not of course include the owners of land or capital). This represent a revolutionary change, completely unforeseen by Marx, the champion of labor, or indeed any present economic analysis.

This vast empowerment of capitalists and landlords over labor carries explosive socio-economic ramifications and produce many divergent scenarios. It opens up for instance the specter of a true free market society where all are capitalists, or perhaps a true command economy where one person or group of persons control the production of almost all goods and services. We look in other words at extreme egalitarianism or extreme totalitarianism, and no present economic theory incorporates such scenarios.

Some may of course argue that these transhumanist technologies may not be carried far enough for such dramatic socio-economic changes. But this is probably an ostrich reaction. Take the much maligned AI for instance. A lot have claimed that robots may never take over our labor force because human work is unique and irreplaceable etc. For one thing, the industrial revolution has shown how better and unique work by craftsmen have given way to standardisation and mass production by laborers specialising in merely one part of the whole work. Economics and mere cheapness can easily override the need for quality and the 'human touch'.

For another, the future dominance of AI can be ascertained with basically one question: Can human minds only function algorithmically? If yes, that's it. Computers too are algorithmic machines, and by all appearance, potentially much better ones. If not, much of our work are still fundamentally algorithmic and therefore AI-replaceable. We cannot so easily escape the specter of widespread human obsolescence. And our economic and political systems are simply not adequate to deal with such levels of unemployment. A super-welfare state where robots work and humans are the bosses and parasites perhaps? But is that really a satisfactory solution?

With genetic engineering, we open up the possiblity of vast inequality on the fundamental biological level (not merely socio-economic, as in AI). Since the service of genetic enhancement (whether at the foetal level or not) is an economic transaction, the question of wealth distribution becomes very pertinent. How do we make sure this enhancement is affordable to all? Or SHOULD we? Also, what about those who refuse enhancements for personal reasons, how should we deal with them? What about those less advanced nations? Will there be an insidious genetic gap? All these become pressing questions.

With cybernetics, i.e. the true systemic integration of man and machine, we face the same problems. In fact, it is probably even more revolutionary and difficult to manage than both AI and genetics, for essentially, it uses one to extends the other, i.e. AI technologies are used to further enhance genetic modified beings. We face certain insidious problems--the greatest of which is probably the simple definition of what constitute 'humanity' and the corresponding assaults on our 'intrinsic value'.

For cybernetics, once perfected, actually allows us the creation of a 'superior' and possibly inhuman or posthuman race (depending on how you look at it). Indeed, it is perhaps the general tendency of transhumanist technologies to be able to vastly exacerbate the inequalities of society. The owners of AI and capital will be powerful enough, but if genetic engineering can improve their DNA and biological capabilities, and cybernetics can turn them into 'supermen', the economic and political power of this elite will easily dwarf the social dominance of even the Nazi or Stalinist elites. A true and supreme aristocracy is thus a distinct transhumanist possiblity.

And how should this elite treats the lower ranks? As truly inferior beings? As being 'merely human'? Or could the whole of humanity ascend together in technological nirvana? Present social, political and economic conditions forbids it. And even if it truly could be done, is it necessarily a good thing?

Worsening the general stew, as Bill Joy wrote, AI, genetic engineering and nanotech (a key ingredient for all transhumanist techs) are KMD technologies (knowledge-enabled mass destruction), i.e. vast violence at a minimum of effort. Even laypeople could do it (witness the proliferation of computer viruses for instance). Such dangers in the hands of the masses gives a huge boost to a demand for a society that is safe at the cost of freedom. Hence again, the totalitarian scenarios. And even if mass terrorism and political repression can be averted, legitimate national states are endowed with unbelievable military and economic power by the transhumanist technologies. Could a simple arms race cause us to give up our humanity? We have done so for much less.

With all these specters in mind, it should be clear I think, to liberals, humanists and general religious that there is a total and extreme necessity for new forms of economic and political systems, should we wish to preserve some of our distinctly 'human' values in the coming era, where for the first time, we can modify the hitherto recalcitrant 'human nature'. There are far more questions than answers currently, but there is still time, for we are still in the earliest stages of the revolution.

After all nightmare scenarios are not the only possible outcomes of transhumanist technologies. As mentioned before, AI allows us to go to the extremes of egalitarianism or totalitarianism. It also frees humans from most physical drudgery and allows the possiblity of distinctly 'human' work (what this may be will be considered elsewhere). AI may also prove to be a superb addition to our general intellectual resources. Genetic engineering on the other hand allows us potentially to elevate the gene pool and biological functioning of the whole human race to a new plateau. It also allows us (in union with nanotech and ecology and maybe undeveloped systems sciences) to manage the biosphere and allow it to flourish in a way hitherto unseen. Even cybernetics (despite its countless dangers) can be turned to general economic and social benefit. All in all, transhumanism is also perhaps the main path towards enabling a true space-faring civilization. That alone can justify a lot.

The choices are not so much technological but political and economic. We should start thinking about them, and start now.

No comments: