This is part 2 of a 5 part series. You can read Part 1 here.

About the Artificial and Unnatural

It is important to take a step back and think about a few distinctions we are fond of making that may not be as useful as they once were. When I published a description of the 2009 talk these posts are based on, I used the phrase “alternative intelligence” instead of the far more common “artificial intelligence.” This is because I do not believe that the distinction between natural and artificial is useful when it comes to intelligent technologies. I also do not believe that the distinction between natural and unnatural is useful most of the time.

Everything we have knowledge of and everything we create is part of the same universal system obeying the same universal laws with results that cannot in any way be determined to be unnatural or artificial. We certainly can have causes and effects that are unpleasant from our perspective. The possibility of that is part of what I am talking about in this series of posts. However, it is not accurate to think of them as unnatural or artificial. Any result we see or can produce is a result of what is possible in the universe, and thus, a part of nature.

It is important to make this distinction because, accuracy aside, I believe that humankind has indulged a myth of the separation of human endeavor and production from the constructions of nature to our own great confusion and detriment. In this way, we justify acts of incredible violence within nature and mollify ourselves about the potential consequences of our technological progress. So let’s be frank and honest. Alternative intelligence of superior stature to our own, should it come about, will be an entirely natural extension of, evolution of, intelligence on this planet and in the universe.

When I begin to remove these distinctions and view these developments as part of a continuum, certain things start to make more sense. For example, let me extend the idea of “alternative” intelligence to include the idea of “alternative vessels of intelligence.” Until not so long ago, I was enamored of the idea of human space travel. I’ve even done a couple of peer-reviewed papers on the subject and worked on design proposals for the interiors of the International Space Station. More recently, though, I have lost my enthusiasm for human space exploration, largely because I cannot figure out where there is for flesh and blood to go. There is no destination reachable within a current human life span that is hospitable, as far as I know. There may well be earth like planets elsewhere in the universe, but we are walled off from them by distance and the time it would take to travel that distance, unless we find some version of Star Trek warp drive. Space tourism, manufacturing and mining is the best future I can paint for humans in space at the moment.

Far more reasonable and likely to me, based on my limited knowledge of what is going on, is that completely alternative forms of intelligence will do the work of exploring the solar system and beyond. It makes much more sense to design vessels of intelligence that are suited to the environments in which they will be placed.

I suppose this could be a highly engineered version of human flesh and blood, as imagined in the movie Blade Runner. Ray Kurtzweil, for example, believes that our robotic technologies will begin to merge with our bodies, with a complete merger scheduled for the end of this century.

Medical science has been replacing parts of us with engineered alternatives for some time now. However, the circumstances under which this technology is being deployed are shifting. Individuals are starting to tailor their bodies through surgery to gain a competitive advantage over their fellow humans. For example, special ops military personnel are surgically enhancing their vision to be better than 20/20.¹Again, there is nothing too surprising here except when you begin to extend the implications of all this engineering to its logical conclusions.

Computer scientists project that by 2020 we will achieve a computational device with a capacity that is equivalent to the human brain. By 2025, they say, such a device will be available for our home office.² Such an achievement would not be human like intelligence, but it is the next threshold we pass on the way to an intelligent being composed of something other than flesh and blood. Hans Moravec projects such a being by the middle of this century.

Of course, the information super highway is littered with the road kill of prognostications and prognosticators who have been wide of the mark. It is worth noting, though, that this is true both in terms of overly optimistic projections and unduly pessimistic ones. Moravec himself describes in detail the painfully slow development of a technology that can drive a car down a road without human assistance. Way back in the 60’s he and his colleagues felt it should be possible in the near term to create such technology. What they learned is that one of the things the human mind is really good at, spatial perception and the ability to distinguish what is important to the task at hand from what is not, is, or at least was, incredibly difficult to replicate in machine intelligence.

It has taken over 40 years to arrive at a place where we are beginning to hear about tests of practical vehicles that will navigate highways by themselves. Indeed, we already have vehicles on the market that can park themselves. What needed to happen is the exponential increase in computational power that we have experienced during the last 20 years.

What we are beginning to have is all the bits and pieces of a new kind of perceptive, mobile and interactive intelligence. Where is all this heading?

This is the end of part 2 of a 5 part series on AI. Next week I will discuss the question of where all this could be heading. Subscribe here to make sure you don’t miss an installment.