The Creeping Revolution or... Meet the New Boss, Scrap the Old Boss -- AI, Soc, $$$

*
Well, if you enjoy the self-assurance of knowing, whatever the AI enthusiasts may say, that artificial intelligence research is still limping aimlessly along at a snail's pace -- and possibly in the wrong direction... well, if you prefer to think that way, please skip this article. And also this one, which discusses another incremental step towards rudimentary intelligence in machines. The mobile robot system IFOMIND is described thus:

IFOMIND reacts initially in an 'instinctive' way to its first perception of an unknown object that it encounters; at first it is generally fearful. However, because the robot is equipped with a human-like capability of inquisitiveness, it realises that it can react in another way and does not have to be scared of something that may not be harmful to it.

So, the robot observes the object from a distance and takes note of how it behaves and how it reacts to different approaches; the robot is then able to decide on the best approach or whether to avoid it. This knowledge can then be retained by the robot as it carries on and meets many more objects.

This is a leap forward in the domain of Machine Intelligence as IFOMIND is able to use logical thought processes in order to decide the best way to interact with the objects that it meets. David Bell from the IFOMIND team explains, "A system that can observe events in an unknown scenario, learn and participate as a child would is a major challenge in AI. We have not achieved this, but we think we've made a small advance."

Frankly, I "got nervous" when non-AI computers evidenced the ability to perform scientific research -- in one case by sifting through medical reports looking for potential dual-use/multi-use drugs, and in the other by systematically testing the genome of a simple animal one gene at a time to determine what each gene does. That was pretty stellar.

But this kind of modest, steady progress can be revolutionary also -- even if it doesn't lead to omniscient god-computers or rebelling robot armies. Why? Because these incremental steps towards adaptable systems could eventually lead to the development of machines capable of independent, relatively unsupervised action.

Imagine you've got an automated Taco Bell. But instead of just having a robotic assembly line churning out your food as soon as you've punched in your order and paid at the drive through, imagine that the "owner" isn't human either. Instead, a limited AI did an analysis of traffic flows and consumer preferences in your area and bought this particular franchise based upon its own best projections. It may have hired a part-time human manager or technician to keep an eye on the business, or it may have been able to replace those positions as well.

Does buying and managing a Taco Bell seem like a complex series of tasks? Then consider an even simpler proposition. Suppose you have a computer analyzing data about the movement of consumers and general foot traffic in a mid-sized city. Said computer then determines where the best locations are for locating vending machines, based upon a relatively simple equation that takes into account the cost of renting space, known buying habits in established locations, any demographic information available, and, of course, the cost of actually relocating a vending machine, so it doesn't move them around willy-nilly.

Said computer, operating through email, post, voicemail and the occasional "face man" human representative, could essentially be "in charge of" an entire business operation for that city -- all the Coke machines, for example. Once the model worked for one city, and one brand of vending machines, it could be extended to others.

Why would this be revolutionary? Simple businesses like fast food, holding rental properties or even managing vending machines may not be very complicated, but if they can be run by computers with virtually no human oversight -- even with regards to critical decision-making -- then you have essentially introduced new actors to your overall economy. Instead of merely attempting to enhance the productivity of our exisiting workers for a particular business, we may be able to create businesses that operate almost independently of human oversight. In other words, we may be able to take a step towards that mythical notion, the "Living Corporation."

In particular, those businesses which can be broken down into basic mathematical formulas and simple elements (basic products, agreements, locations) with few variables (such as the vending machine scenario above) will be particularly workable options. Instead of being "cogs in the machine," these machines will be driving economic decisions. Now, you might ask, "How does this differ from the automation that's been going on since the Industrial revolution?"

This situation differs because the historical trend in automation has generally been to replace unskilled workers engaged in brute labor with more skilled and/or educated workers. So you might substitute one semi-skilled worker with a bulldozer for 50 unskilled people with shovels. Even when people doing mentally challenging work have been replaced by machines -- such as accountants being replaced by user-friendly accounting programs -- the trend has been to eliminate only repetitive tasks requiring little judgement.

In theory, the above machines would be following this trend, except that we would no longer be retaining that last worker or workers. The machines, to an extent, would be capable of looking after themselves. Which means that instead of assisting those who drove actual economic activity, they would be driving that activity themselves.

You might also ask, "How does this differ from computers which are doing independent biotech research? How is replacing entrepreneurs and managers a dramatic improvement over replacing scientists?"

Well, in many ways the digital businessentity and the independent cyberscientist are co-equal as revolutionary forces -- because they fill the same role. You're not just increasing worker productivity, you're replacing critical workers whose efforts are key to driving the economy. Instead of just trying to increase the productivity of a certain set, finite number of workers in your population, you are effectively increasing the size of your working population altogether. And you are increasing that population without expanding the overall demand on most of your infrastructure. (Electrical and communications grids, yes; housing and transportation, no.)

A common concern in U.S. high-tech executive circles is the lack of new American engineers and scientists being produced by American universities and the faltering stream of technically immigrants compensating for that shortfall. If computer systems can be designed to replace multiple scientists and technicians in doing critical, if basic, scientific research, then it is possible to dramatically increase the number of de facto scientists in your country by acquiring and utilizing vast numbers of such systems.

As with the "businessentity" programs, some science will be easier to adapt to automated researchers. The Human Genome Project probably contained some early examples, biotech research could undoubtedly make use of more (if only by expanding upon the two systems already active) as could any field in which the systematic collection, parsing and coelation/analysis of data would be profitable. For example, cataloging every object in the visible sky while noting anomalies, likely planet-bearing systems, etc. The most complex and inspirational kinds of scientific work like developing major new theories will still require human scientists. But this latter work will only require the best scientists we have today. People who are doing plodding, pedestrian work -- or who are doing a lot of repetitive "grunt work" for a paycheck or as graduate students -- will be in far less demand. Our ability to replace or even multiply less brilliant scientists with computers changes the entire research and development end of a region's economic equation.

How immanent are the "living corporation" programs in this equation? Sadly, there's probably a spambot program out there which already meets these criteria, and probably more than one currency speculation system which does so. These, obviously, are fields where the business was easily broken down into numbers, equations and discrete products and strategies. But as the technology and our understanding of its potential improve, more and more status quo businesses will be interpreted into a virtual battlefield which these machines can understand and thrive in.

How will these changes effect economies and societies? I strongly suspect we will find ourselves in a race with our own machines as well as the rest of the world. But it is not necessarily a race we don't want to run. If artificial entrepreneurs and pseudo-AI scientists become widespread, and if, at the same time, more and more manufacturing and physical labor are being automated as well, we may discover our strongest economic and human imperative is to educate people to fill those many jobs our robots and computers will still be unable to handle.

Using various methods of human augmentation, creating superhumanly intelligent and creative scientists, entrepreneurs, artists, writers, leaders, etc may prove to be relatively easy. The will to embrace any form of radically improved human ability may prove very challenging. But faced with the alternative, many people will take that option, rather than be left in the dust.


Future Imperative