I am doing the IB diploma, to complete this I have to write to also write a theory of knowledge essay (1500 words max), for this essay I chose the topic : Can A Machine Know?
I have releveant examples from both sides of the discussion but for my first draft I chose that No a machine cannot know?
I am posting my first draft here so if any of you who have time can give it a quick read and let me know what you think (english is not my first language).
Also it would be great if you could tell me what is good and what should be taken out, also give me your opinions on this discussion along with examples (include links to them if you found them on the net or books and stuff).
Thanks and here it is.
Can A Machine Know?
Can a machine know? This question has grabbed the attention of philosophers and scientists alike for centuries. There is still no concrete answer so far but there are numerous takes on this, there are opinions which state that a machine can know and there are equal numbers of opinions which state that a machine cannot know.
This so called “knowing” capability of a machine no matter if it exists or not has been slowly replaced by the allusive term “Artificial Intelligence”. The concept behind this is completely logical but weather it exits or can be produced poses a great deal of confusion.
To be able to answer this question we must first set laws and rules of what “knowing” really is so that we can test if machines pass these rules. So far “knowing” has been classified as the ability to comprehend and reproduce the subject in consideration provided that it has been experienced before.
There are many strong arguments and reasons that can lead us to think that yes a machine can know.
One such example is the chess playing machine named “Deep Blue”, of which the creators stated that it can beat humans at chess, this seemed hard to believe at first but when put to the test it exceeded the expectations and defeated the worlds best chess player. This event immediately triggered the mind of philosophers since chess had always been considered a hallmark of human logic and reasoning. All the moves in chess were fed to this machine via programming, now doesn’t that follow in the law of knowledge?
Another example is the robot that preformed unassisted heart surgery. The explanation provided for this event by doctors is that this robot had been involved in many assisted surgeries and that it had “learnt” from them. This robot was able to do this via experience; doesn’t that also fall under the laws of knowledge?
Now what if we combined man and machine, no one has ever said that A.I cannot exist in that form. The United States government is planning to do exactly this; they intend to turn soldiers into super soldiers by fitting them with powered exoskeletons. This technology will help soldiers leap extraordinary heights, expand their vision by a factor of 3 and amazingly the exoskeleton will automatically perform necessary first aid maneuvers on the bearer if needed. These new bread of soldiers are expected to be unleashed in the coming 100 years. This falls into A.I since the suit is able to perform first aid maneuvers automatically without any help from administrators afar.
Another major example is computers; they have revolutionized the way we live, we have learnt to trust them enough that we every day trust the live of numerous humans with them. Computers know since they always seem to give the right answer in many computations, they also seem to follow trends for example If I write one line in red, the second in blue, the third in black and then repeat the cycle in a text document, the computer will automatically select the color I want in each line for me.
Simple amazing but…
There is an even stronger reason to believe that a machine can not know.
So you say that if the chess playing machine “Deep Blue” can unseat the worlds best chess player, when will a machine unseat us? The answer in my opinion is no unless in the future the cerebral capabilities of humans can be produced in machines.
So what if Deep Blue can overtake the world’s best chess player, can it overtake the world’s best Backgammon player? No it can not in fact it is incapable of doing literally nothing but playing chess. Even if we take this machine and show it how to play backgammon through different means such as reading, talking and showing it to it, it will still not be able to do anything but play chess since machines so far cannot absorb new knowledge automatically like we humans can, and this does not fall under the laws of knowledge.
The same applies to all machines invented so far, they cannot perform anything else than what they have been programmed to. Today every one has used a calculator and we all think that it always has the right answer, does it? Like every machine up to date, even the most powerful calculator has been programmed by a human, what if the human who programmed a specific calculator by mistake so that the output it would give to 1+1 would be 4, will the calculator be able to understand by it self that 1+1 is equal to 2? No since unlike humans machines lack the ability to use logic and reasoning.
As for the robot that has successfully carried out heart surgery based upon past experience. What if the surgery to be preformed was slightly different than the ones it had learnt, would it still have been able to successfully perform it or not? The answer, we can never know since the event did not occur but knowing that a professional doctor was observing the surgery while sitting at the other end of the world gives us a clue, that it probably could not have, this was done because machines do not posses the ability to use logic and apply it to different situations and that is why we can never trust machines blindly, there always has to be a human at the end of things, in other words there is always room for error with machines.
Now let’s take a computer, we say that it knows how to perform tasks and can even recognize text as voice and read out text written by a human. This is not entirely true since a computer can only perform an indicated task if the human using it gives the computer the proper command associated with that task string. For example in the “Backdraft” distribution of Linux the command to obtain the boot key of a windows system is “bkhive /mnt/sda1/WINDOWS/system 32/config/system key” but in a different distribution of Linux the command for the same task is completely different.
Also when inputting this command the user has to tell the computer the exact path of where the boot key is located; the computer is unable to find it itself.
As for the ability of a computer to read out text, it is true it can do that but it does so in a single and non articulating tone; it is also incapable of recognizing punctuation used!
Now lets have a look at organization, a human can organize clothes in a wardrobe in a desired way if shown before, but when we place files in a computer a certain way; no matter how many times we place them in that way the computer will be unable to organize them itself.
So therefore in conclusion my answer to the question “Can a machine know”, I choose to answer no, it cannot know. My reason for this is that machines do not posses the ability to acquire more knowledge like humans do; machines cannot use logic and reasoning.
I believe that when people say that a machine has learnt something through experience, it is fake since once programmed the machine cannot learn anything new unless the program is updated, that is the exact reason why many technological companies frequently keep providing updates to their products since it is humans that continuously feed machines with knowledge. Also I believe that if there is no way found to replicate the exact cerebral capabilities of humans in machines, a machine can never know.