Archive

Archive for May, 2011

Dumb Phones and Artificial Stupidity

May 29, 2011 1 comment

The phrase “artificial intelligence” is a misnomer because what is actually being imitated by all A.I. programs is not what humans call intelligence, but rather what we tend to refer to as stupidity: that is, doing something without thinking. The common thread among computer intelligence programs is to turn an activity that seems to require intelligence, or once required intelligence, into a task that does not; to tear it apart and strip it down by turning it into something that, well, a machine could do. In doing so one shows that the task actually requires no thought. In this way, artificial intelligence does not imitate intelligence, but rather demonstrates that no intelligence is actually required to perform the task at hand–provided one has adequate instructions. This dawned on me clearer than ever as I played a computer in a “rock, paper, scissors” simulation. The simulation revealed to me not how smart the computer was but, rather, how dumb of a game rock paper scissors actually is. The fact that the game was “rock, paper, scissors” was key to my epiphany, since I already knew that this was a dumb game, and thus that the computer was pretty dumb for…taking it so seriously.

Let us, however, consider the more famous, and relevant, examples of artificial intelligence. The first is the computer victory of “Deep Blue” over chess champion Gary Kasparov. And actually, I think I have been too harsh in not mentioning yet what is impressive about artificial intelligence, in this case and in others. What is impressive is the programming task, which is distinct from the program itself. Programming is distinct from the program created in the same way that an artist is distinct from his creation. And computer programming, believe it or not, is a sort of art, in the same way that mathematics is (computer programming is basically just a special applied kind of mathematics). I was impressed by what the programmers of the Deep Blue program actually did. Their accomplishment was impressive because chess has long been believed to be the ultimate strategy game, requiring intuition, planning, and creative thinking on the fly.

But the team at IBM cracked chess–somewhat at least. They demonstrated that there is an algorithm for an extremely strong chess strategy. This is an amazing intellectual accomplishment. I am honestly amazed by it, because I wouldn’t have guessed it was possible. But it is the programming achievement that amazes me, the cleverness of that, the thought that went into figuring out how to create a winning algorithm. If you do some investigating, you will find out just how much thought, and research, went into this. It is truly impressive. But we must remember not to mislabel things: the algorithm, the final program that is the product of the programmers efforts, is not smart. Nor is the machine that implements it (Deep Blue). Rather, it is the people who wrote the algorithm that are smart. It is their intelligence that is on display, not that of any machine.

When you think of it like this, you will realize what a farce the concept of artificial intelligence really is. To whatever extent we think that we see intelligence in computers, we are seeing the intelligence of the programmers reflected. Sometimes it is buried so deep, and is so complex, that we think the computers are smart. But this is an illusion, on the same level as the magician’s illusion: when you see how it’s really done, the magic goes away. You will realize that the laws of nature have not been violated, after all. You may still be impressed with the magician, however. In the same way, when you get a basic idea of what goes into A.I. programs, you will see a mind at work in the program that was created, but you will also realize that there is nothing like a mind at work in the program itself.

More recently, we have the computer program that was able to beat two Jeopardy champions. Again, watching this, I was struck by the depth of the accomplishment: that the programmers were able to cover most cases of what the format for a jeopardy clue tends to be, and to write a program that quickly accessed a database of information in such a way as to make a reasonable estimate as to the answer (actually, the “question,” in Jeopardy terminology). It is fun to think about how they did this; to look at a jeopardy clue and imagine the sort of things that the algorithm looks for–key words and phrases, etc. And in doing so you realize that Jeopardy clues, with their multiple layers of hints within the clue, are especially amenable to the sort of analysis that one can program a computer to do. I was impressed that the programmers accomplished putting this together in a way that was able to handle most clues and didn’t mess up too often. But I was not impressed by “Watson” himself (nor, by the way, was I impressed with his humongous buzzer advantage, which was glossed over by the media coverage of this publicity stunt).

The real problem is that we do not, today, have a good idea of what intelligence is–and this stems from the deeper problem that we have a mechanistic concept of the mind. We tend to associate intelligence with calculation and precision, which are two things that computers are great at, two things they have been better at than humans since their invention. Computers will continue to grow faster, as time goes on, and their programming will become increasingly sophisticated. And if we continue to misunderstand intelligence, then we will think that they are progressing towards us in it. The truth is that computers have always surpassed us in doing many things that were never a mark of intelligence at all, and meanwhile have not moved an inch closer to becoming something that they intrinsically are not. Processing speed has never been the issue with computers being intelligent. The real issue is that a computer is nothing like a mind. The similarities are all superficial, and all based on analogies that have, over the years, seeped into the public consciousness. The analogies go both ways–we say, for example, that computers have “memory” and that human beings can be “wired” to act a certain way.

What I am advocating is an attempt to change the way we look at computers. Do not ever approach a computer with the idea that it is anything like a mind, and you will never be in danger of thinking that it is. Everyone’s first intuition about all computers is that they are dumb. And this is the correct intuition. Dumb as a rock. It is repeatedly confirmed by experience. Computers are exceedingly dumb–they are not capable of even the simplest reflection, of any measure of creativity, of any measure of wisdom, or anything else that smacks of any sort of awareness of reality or self whatsoever. Because computers have no awareness. Moreover, they never will; no matter how fast they get, no matter how advanced their code is. Why? It is obvious. They are machines that implement algorithms. We can all see this. And we can all see (despite what science fiction tells us) that we are not.

Our minds, and even our brains (if you believe there is a distinction) are not computers, and are not like computers. I’m not sure what they are, but they are not like computers. One problem, in realizing this, is that there are many very smart people who think this observation is up for debate, and will engage you in some intractably difficult philosophical discourse if you suggest otherwise. They will ask you: do we really know what a mind is? That it is actually distinct from matter? Isn’t the brain just a collection of atoms interacting in a complicated way? Well, I’m not sure–and I think the burden of proof is on the materialist, and not the other way around. But I think this is all quite beside the point. The point is that computers are not like minds, even if they are just a collection of atoms; we might as well believe that someday we could shape a potato (another collection of atoms) into something like a mind. It is on the same level of absurdity. It is as absurd as the ancient worship of idols:

Their idols are silver and gold,

They have mouths, but they speak not;

They have ears, but they hear not;

They have hands, but they handle not;

They that make them shall be like unto them…

We scoff today at the idea that ancient people could believe that a metal idol could really be a god with any kind of intelligence. But this is only because we find different things persuasive. To them, the convincing face may have been enough, and they disregarded the obvious (to us) fact that it was just a hunk of metal, created by men. Today we see computers being designed to imitate something we more naturally worship–reason and logical deduction–and we jump to similar conclusions.

The most destructive idea that comes out of all of this is the belief that technology, the fruit of the computer revolution, because it is smart, will make you, and us, smart. Listen to advertising today. “We are building a smarter planet.” What? They want us to believe that “technology is smart and will make the world smarter and will bring it together.” This is not true. In fact, technology is very dumb, and is making us dumber, and will bring chaos and misery to the world…if we don’t handle it properly.

Technology, handled properly, can be good for us, or at least neutral. But the more we, without reflection on what we are doing, integrate it in our lives, the less intelligent we will become, the more disconnected from each other we will become. And why? Because we will become more like computers, which are dumb, and have no interest in reflection or beauty or creativity or community or anything else even remotely human.

The irony in all of this is that the gap between man and machine is most at risk of being bridged the other way. If our culture continues to push the analogy between man and machine, this will do nothing to make them more like us, but it will do quite a bit to make us more like them. So please, challenge the language of today, in regards to the computers that are becoming increasing involved in our lives. You don’t have a smart phone. You have a dumb phone. There are no smart phones. That urge you get to scream at your phone, to tell it that it is a “dumb phone!”: that is the correct urge because that is what the gadget is it is. Your instincts are correct. Your phone is dumb. Also, your ipad is dumb, your desktop computer is dumb, the GPS system in your car is dumb, and the internet is dumb. This laptop I am typing on is dumb. The people who created these things, however, are mostly all geniuses, and much credit is deserved for their accomplishments. But, all the same, I do wish they’d give it a rest. The technological revolution is truly exhausting.

Advertisements
Categories: Uncategorized