Kayson Wong
I love computers. Of the many reasons, one of them is that computers do not lie. It’s true. Computers are machines, they do not think, they have no feelings. Computers always give you the exact same output when given the exact same input. In fact, it’s all logic.
What’s logic? Logic is something that everybody on this world agrees. You must agree, and you have to agree. That’s how logic is defined. When you ask everybody the same question, they always give you the same answer. Mathematics is a field that is derived from logic rules. For example, if you ask someone “What is 1+1”, they always reply “2”. It’s not because they think or feel the answer is “2”, it is because they’re told that the answer has to be “2”. By the definition of numbers and operations, “2” is the only answer to this question. Computers are designed with this same idea. At the core of a computer, logic operations are performed by the CPU. Computers will always produce the exact same output if they’re given the exact same input. This is the characteristic of any computers on this planet.
Today, many computer scientists are trying to push the limit of our computers further – to something so called “Artificial Intelligence”. Artificial Intelligence, or AI, is a term used to describe algorithms that simulate tasks that are normally accomplished by human beings only. As a result, decisions that were human-made in the past are now handled by computers, or to be more exact, the programmers that came up with the algorithm. Some people today like to refer to this as “smart” technology. Camera manufacturers design AI algorithms so their cameras can “smartly” determine the “appropriate” shutter speed, aperture, focus and ISO speed for taking a photo. Translation software can translate an entire paragraph of text from one language to another. Robots can play table tennis. However, many AI algorithms are applied inappropriately, and in a lot of cases just make the problem worse. Professional photographers never use auto modes to take pictures; in fact professional cameras never include an auto mode, simply because the shoot settings calculated by the AI are far from the ideal settings for taking a photo. Language translators, up to this day, often distort the original meaning of the text, or even come up with grammatically weird sentences. As for the robots, I cannot understand why someone would design a robot to play table tennis. These ball games are designed for human beings for fun and health, not for the sake of ball games.
Putting computers into the seat of decision making is even worse when safety is a concern. In Netherlands, the Maeslantkering is a floodgate designed to protect the city from storm surges. The gates are entirely run by a computer; no human intervention can override the computer’s decision. The computer is programmed to close the gates if the surge height predicted by weather models is higher than 3 meters. During a storm, it happened that the computer predicted the surge height to be 2.99 meters, so it left the gates wide open. Obviously, no one living in the city would care if the water level is 1cm higher, but it matters to the computer!
Human beings must be involved in decision making processes, because unlike computers, humans can react to the situation creatively and come up with solutions that best resolve the problem. Despite the technology we have today, pilots remain essential in an airplane, because we believe that shall an emergency occur, pilots can react creatively and come up with the best way to solve the problem. Engineers cannot foresee each and every possible situation that can happen, so we need human beings in the cockpit to make decisions, despite of the fact that sometimes they make mistakes.
Another decision-making process that is more abstract and convoluted is the creation of arts. While some may argue that in the future computers can accomplish this job, they cannot. It’s not because of technical difficulties, it’s because this task is theoretically impossible to be achieved. Recall that computers are operated based on logic, every computer will give us the exact same response when the conditions we input to the computer are exactly the same. Can we convert the process of creating artistic materials into a logical process? No. Every one of us has a different feeling towards arts. I may look at a painting in a museum and see nothing; the guy next to me sees everything. We all interpret arts in a distinct way, a way that is based on things we have seen, the skills and talent that we have, our personal characters, and even our mood at that time. If we ask two persons to write some music, it’s very likely (if not for certain) that the two pieces of music will be completely different. Every person in this world gives us a different piece of music, because our DNAs are different, our experiences are also different. All the computers in this world, on the other hand, are always the same. This fundamental disparity distinguishes human beings from computers. So, it is not possible to come up with an algorithm that can replace human beings in art creation processes.
But what if we’re not trying to come up with an algorithm to write all music in this world, we’re trying to ask the computer to write the music that Beethoven himself would have wrote? If we are to simulate such creation process by a computer, we must answer the question of whether a human being is “computable”. That is, can be write an algorithm to simulate a human being?
The answer to this question is also no. To simulate a human being, we would have to simulate every physical and chemical change that occurs inside a human body. This requires simulation at atomic level. If we can calculate how each and every atom in a human body would behave, we can then work out all the chemical and physical reactions, and thus, simulate the human body. The truth is we cannot simulate every atom, because it is impossible to predict how a certain atom would behave in a given situation. It’s not even possible to measure the state of each atom. There is a theory in quantum mechanics called “uncertainly principle”. The short version of this principle is: if you measure the position of an atom to very high accuracy, then you would have a very low accuracy about its position, and vice versa. That is to say, we cannot theoretically take absolute accurate measurements; there are always errors to a certain extent. Due to this error, we cannot exactly predict how each atom would behave. Quantum mechanics actually is a study of probability of atoms. For example, there’s 70% probability that the atom would pass through a certain region, and there’s 30% probability that the atom would be bounced back. It’s like throwing a dice, nobody knows the result. Fortunately, this probability is very useful because usually we’re dealing with a very large number of atoms when applying quantum mechanics. If we throw 6 million dices, then we can be quite sure that about one million of them will have the face “1” on the top. As we reduce the number of dices, this one-sixth prediction become less and less accurate. When it comes to one dice, it’s very hard to tell the result. The same applies to the atom case. This is why it is impossible to have physical simulations up to atomic level. Even if we can, we have to ask another question: is a human being a “computer”? That is, if I have two persons who are exactly identical, who share the exact same DNA and have the exact same experience, will they react exactly the same to a certain situation? I doubt it.
So at the end, what are computers good for, if they can never be as smart as humans? The answer is repetitive tasks. A lot of our economic activities involve mass production, which means producing a large amount of the same product to minimize cost. In the past, without technology, human workers are hired to this job. They have to repeat the same tasks over and over again, every day, every week, and every year. Unfortunately, human beings get bored at repetitive stuffs. As is turns out, these human workers do not enjoy much of their life. Today, these jobs are replaced with machines in factories. Machines can accomplish the same task over and over again, and more importantly, machines never get tired, get sick, get bored, or make errors. Machines are ideal for this job.
The decision making part is the part that should be left to humans. We’re not just living in this world for the sake of living; we’re living because we want to enjoy life. If the machines can create music, make paintings, feed us and talk to each other, why do we need to live? We’re not making computers into humans; we’re trying to make them help humans. Machines are merely the tools we use to accomplish our goals. In this context, it’s best to let computers or machines take over the dirty, dangerous or boring jobs on this planet, and let humans enjoy the sunshine at our beautiful beaches.
References
- Borel, Brooke. A Ping-Pong-Playing Terminator. 2 16, 2010. http://www.popsci.com/technology/article/2010-02/ping-pong-playing-terminator.
- "Maeslantkering." Wikipedia. n.d. http://en.wikipedia.org/wiki/Maeslantkering.
No comments:
Post a Comment