There’s been a lot in the news lately about artificial intelligence and robotics. They just broadcast a PBS “Nova” on the topic, and it was a featured topic of discussion on NPR’s “Science Friday” today. People are starting to notice that a lot of jobs are threatened by AI, and not just drudgery-type jobs like in manufacturing and agriculture, but in management and engineering, even the arts. Cabbies and truckers are already on the hit list, how long before airline pilots and ship captains get pensioned off? Sure, technology and automation have always changed human employment, made mindless mechanical and repetitive skills obsolete. But this is different. Now we’re talking mental skills that used to require human judgement and intelligence. Steam engines made bosuns, topmen, riggers and deck apes obsolete, but processors and satellites made sextant navigators redundant. It feels a little different when the job-killing machines shift their attention from the shop floor and start savaging the front office. Suddenly, the technocrats and administrators are finding themselves replaced. Yeah, yeah, we’ll still need people to program the machines, but we can’t ALL be coders, can we? Even the military has felt the pinch. We don’t need millions of infantry any more, the fighting and dying is now being done by brigade and batallion sized units. Before long it will all be robots and Beserkers, or the wars will be won and lost in cyberspace.
I think this is a very old problem, not a new one. We tend to think of intelligence as being carried out in complex systems, either highly evolved biological nervous systems like ours, or highly engineered and computerized digital systems–the traditional concept of AI. But conceptually, a “brain” or “intelligence” is a highly interconnected collection of discrete components operating as a system. It detects information in its environment, interprets it, and then acts on that environment to bring about a desired result. Our traditional idea is of an array of sensors and a collection of servos, with a central processor between the two to manipulate the data and make decisions. A parallel memory storage device contains both the intermediate results, and the recipe of job instructions (the software). It doesn’t have to be a biological brain, or a silicon and metal machine. There are other alternatives that have the same architecture.
The “brain” could be a national economy, a natural ecosystem (remember Gaia), a corporate hierarchy, a government, a state, a military chain of command. Instead of wires or neurons, the internal transfer of information in the processor can be carried out by verbal orders or stacks of typed memoranda, papyrus scrolls, cuneiform tablets, queipu, telephone calls or emails. The analogues of sensors and servos can be easily conceived as well: Conceptually, they are all I/O devices, processors, memory, cams and solenoids, gears and pulleys. You don’t need silicon to make an artificial intelligence, all you need is people organized into groups.
The Roman Empire is an example. It certainly exhibited behavior that none of its members could have been aware of, or could possibly have wanted. Surely, all of us have at one time worked for an organization that acted as if it was serving its own needs, not those of any of its members. And just like computers, human social collectives frequently fuck up in truly spectacular ways. Sure, large collectives often have human masters, but so do mechanical computing devices. And they both sometimes exhibit behavior their masters did not desire or even anticipate. Companies fail, civilizations collapse, countries implode, and sometimes, they do quite well in spite of the idiocy of their leaders or their programming. And sometimes they form alliances or fight one another. They are born, live, grow, decline and die. They even reproduce, both sexually and by fission and mitosis. To some extent, they are sentient.
I think we’ve had artificial intelligence of a sort ever since we learned agriculture, literacy and moved into cities. We’re just now getting around to electrifying it so that it operates faster than we can, and with less human supervision. That’s what’s really new, we invented administrative computational systems long ago, but we’ve just started mechanizing them. Don’t let the technical virtuosity of the machinery delude you into thinking they are the creations of genius, or that the systems express some startling success of the human mind and creative spirit. They are clever, but not spiritual. And I really think the key to understanding what is really happening in this so-called high tech society of ours is to view our Digital Revolution as merely the systematization and automation of bureaucracy.