26 Mar 2007

The following is an excerpt from an essay written for Rebecca Rowson's IFSM 304 DE class Term I.

Scientists around the world have been focusing on creating machines of more processing power and higher value to humans ever since a computing machine was invented. The focus lay mainly in manufacturing equipment which could take over assembly line work or take on calculations which would take months for the human brain to solve to name a few examples. This focus has been shifting to incorporate other areas in the past 40-50 years; scientists and engineers have been working on constructing machines that are able to "think." More and more are these machines seen in everyday situations such as aiding doctors in evaluation and treatment possibilities, or in data mining, a technique used by various organizations to filter out certain data in order to create patterns that otherwise would go unnoticed. The media lab at MIT is in the process of creating a common sense database called "open mind common sense," and work is constantly being done in producing a natural language processing computer which could "fool" a human, the ultimate goal of scientists in this field. It seems as though the breakthrough of creating a thinking system that is no longer distinguishable from a human brain is getting near; the difference will be that this machine will be able to better accomplish tasks of any sort in comparison to the human brain.

This paper will focus on the creation of artificial intelligence and its use to humankind. If mechanical calculations are better in assessing the value of our next steps than we are ourselves, should we then simply let computers govern our hospitals, businesses, even countries? It will examine whether the human workforce will become obsolete through this science and whether there is a reason for an instance such as the government or the common public to interfere with or prohibit this sort of science. Also will it focus on whether a computer will ever be able to fully "understand" the human brain and where the difference of these two lies if there is one such, in addition to assessing the term "intelligence" and what makes humans have this.

As Robert Sanders explains in his review of a book written by Gene Rochlin, computers have taken over many areas of our daily lives, thereby alienating the human being who is pushed to the side and made to feel useless. Also, by overly depending upon computers and letting technology take over much of our daily routines, human knowledge is lost, and so are their skills and autonomy (Sanders, 1998).

According to the Stanford website about artificial intelligence, the term can be explained as trying to create machines and software which are "intelligent." The major problem hereby is that there is not yet a solid definition of the word "intelligence," but it is simply related to the human ability to solve problems with the help of computations occurring in the brains. So far, machines have been created that are able to come to "intelligent" conclusions in one specific and well-known area of research, but it has not been possible to create a device that is able to learn in the same way as a human is capable of doing (John McCarthy, 2004).

So far, computers might be able to accomplish great calculations and computations, research enormous databases for possible patterns, and spit out results that seem better suitable for the situation than humans can produce in the same amount of time, but it has not been possible to call a computer truly intelligent because the learning process of a natural brain could not yet be deciphered.

Some researchers such as Dr. Michael Rovatsos from the Center for Intelligent Systems and their Applications School of Informatics at the University of Edinburgh Appleton, who discusses AI in an interview in the Khaleej Times, believe that AI might never reach the level of the human mind in all its complexities. He states that most scientists are not so much in search of reaching this specific goal but are more enticed by the knowledge gained during this research, which can be used to "make our lives simpler." He explains that it is not the common researcher's goal to replace humans with machines, but the moral responsibility of preventing this lies in the individual with the knowledge needed to hypothetically do so. Dr. Rovatsos says that through this research, humans will need to find a different approach to ensuring their position as "crown of creation." Humans have used their superior intelligence to mark this spot as their own, but this may no longer be available if computers do reach this level of sophistication. He stresses the fact that even though the human brain may be a poor calculation machine, it incorporates many beautiful features worth much more than being able to compute large numbers (Rovatsos, 2005).

The questions raised here are how humans should react to such science, whether to believe that the goal of reaching human intelligence is near, far, or perhaps not ever reachable, and how to treat such machines, should they ever exist in the future. There is no doubt that computers and the corresponding science have aided humans in an unthinkable number of ways and were mostly welcomed by the common public. Is it therefore ethical of the public to cry out at this point? Who has the right to censor scientists in one area but let them roam in others, and where does the intimate knowledge concerning this matter derive from which would make it possible to bring forth adequate reasons for such a censorship? The view of what the human brain is might be of help here. There are of course many different approaches to this question; where some explain the human brain as simply being a huge machine sending out signals of ones and zeros through millions of nerve axioms analogous to computer systems, others believe that it is a gift of God, certainly not to be copied by humans themselves. If the brain is viewed as merely being a computational machine receiving energy in the form of food, then there should be no concern in the general area of letting scientists create other machines capable of the same computations. Such a view of course does not imply that this lobby needs to be unconcerned about every issue AI raises. For defenders of the view that the brain is God-sent, the issue arises which technology to use and where exactly to draw the line, as AI is not as clearly defined. There are appliances that incorporate bits of AI knowledge that are commonly used, and who would not want to depend on an AI database if it can give a more accurate description and treatment method of their child's illness?

If machines were indeed to make their way into the realms of the human brain"s abilities, how will we need to treat them? Will we need to present them with the same amount of respect and sympathy as we typically bring forth to our fellow human beings? This question raises a more detailed view into why (in general) we treat fellow human beings the way we do. There are two possible reasons discussed here: might it be because we look alike and therefore think we feel alike? Pain is usually unwelcome and tried to be avoided whereas pleasure is the goal; do we therefore feel consideration towards our "look-alikes" and present them with equally as little pain as possible, while awarding them with pleasure as far as possible? Or is it because we believe that any animal with a brain can feel as we do? There might be a deeper thought: that of consciousness. If a creature is conscious, meaning it knows it exists and has a sense of "I", then humans respect this by giving pleasure, leaving out pain as far as possible (we are purposely omitting the few deviations such as torture and the like to ease this task of conclusion-finding). Considering these thoughts, then a machine must either possess a brain, look like a human being to be considered human, be self-conscious, or incorporate two or all three of these aspects. The brain is called the CPU, perhaps working hand in hand with RAM modules, and scientists can most likely establish the resemblance to a human being's appearance as well. The problem remains of giving a machine the ability to possess consciousness, as humans themselves do not know exactly where this ability lies in and how to copy this. Children do not feel entirely self-conscious when they are born and only learn the full scope of this after some months. Does this mean that infants are not to be considered human until they possess this ability? And does this fact also mean that machines can learn this ability just like humans can? These, as well as countless other considerations, lead to the conclusion that there is no easy answer to whether artificial intelligence can be directly compared to human interactions and brain labor.

We may assess whether the usage of such sciences is ethical in our world and will use the Kantian perspective to come to a possible solution. The three major concerns here are universalization, sense of duty as sole motivator, and viewing humans as an "ends" instead of a "means." Scientists more often than not view their work as a moral duty, helping humankind to better their lives. This surely falls under the sense of duty Kant has spoken about and therefore could be viewed as having a moral aspect. By creating machines which aid humans in their daily lives, for instance by creating data mining apparatuses that may give more accurate conclusions about a certain sickness than doctors can, then this can also be viewed as being moral in Kant's sense, as humans are here viewed as an "ends" instead of merely a "means." The question of whether this can be made universal is a more difficult question to answer, as the common opinion probably allows only scientists with moral objectives on aiding humans instead of doing harm to these should be able to work in the field of AI. But who can differentiate between true desire to help and counterfeited "good intensions"? This thought again undermines the prior assessed moral values of the science of AI and we are thrown back to begin at step one.

If there is an instance which should come to a conclusion of whether censorship is appropriate in this field, the search for the three moral imperatives continues. A universalization is almost impossible, as it is not clearly defined where AI begins and regular machine creation ends. Humans are seen as a "means" if the result of AI is viewed as harmful, but because AI surely has many positive aspects, it is not in humanity's best interest to censor this sort of technology and science. The conclusion drawing instance, which may be the government, might feel a moral duty to stop any research in this area, but by doing so creates a hostile environment for scientists and at the same time gives other countries a huge advantage in the same field if it is not forbidden there. An advantage for one country inadvertently produces a disadvantage for another, which again is not in a population's best interest and so cannot be easily expressed as being moral without a doubt.

In conclusion, it remains a very interesting issue with many opinions and thoughts attached, but no definite truism found so far. Science will advance no matter how many restrictions are placed because this it a part of the human mindset, and we can just hope that scientists have been schooled in issues such as ethics as well besides the technical matters. Artificial Intelligence has aided humanity in various ways and has become an important part of our society; we must ensure that it will only be used to aid in bettering this world rather than functioning as a tool for increasing personal wealth. When used wisely, most humans will not lose their role in society but be able to have added time to flourish in more meaningful aspects. The exact ways of ensuring this are so far not known and humanity will have to take on this obstacle when it is faced by it. There is no saying what might change through this new science and how far its tentacles will reach; we might consider asking a data-mining machine for its opinion, thereby entering the vicious circle once again.

References

(2006). What is the singularity?. Retrieved November 10, 2006, from The Singularity Institute for Artificial Intelligence Web site.

A.L.I.C.E. AI foundation, (2006, October). Putting the puzzle pieces together. Retrieved November 10, 2006, from A.L.I.C.E. Artificial Intelligence foundation Web site.

Joshi, Anshuman (2005, June 10). A matter of Artificial Intelligence. Retrieved November 14, 2006, from Khaleej Times Web site.

MIT Media Lab, (2006). What is open mind commonsense?. Retrieved November 9, 2006, from Teaching computers the stuff we all know Web site.

Sanders, Robert (1998). What can happen when computers take over. Retrieved November 14, 2006, from University of California Berkeley Web site.