The Ethics in Artificial Intelligence Systems
- Length: 3636 words (10.4 double-spaced pages)
- Rating: Excellent
Not too long in history were computers invented. In fact, computers were first invented within this very century. Today we can see that computers are extremely helpful in our daily lives. The field in technology has improved so much in just the last few decades. We can see this by examining the technological advancements in computers themselves--such as having more capacities, storing more memories, performing more tasks and even in a less amount of time, etc. By observing the inconceivable progressions, one can see that not only will the field in technology will grow but grow exponentially. Thus, the growth in technological innovations will be promising to society. With this in mind, for the past few decades, many people especially scientists, researchers, and inventors have devoted so much time and energy in AI (artificial intelligence). The subject matter of artificial intelligence has brought many anticipation in society. Not only does artificial intelligence has many promises such as efficiently solving "numerous technical problems," but also AI would promise for a better understanding in "cognitive processes, particularly, the human mind" 1. Before we go any further, let's take a look at the history behind artificial intelligence.
The concepts of the development of artificial intelligence can be traced "as far back as ancient Greece." 2 Even something as small as the abacus has in somehow led to the idea of artificial intelligence. However, one of the biggest breakthroughs in the area of AI is when computers were invented. So who was this genius that came up with this idea of computers? One would have thought that computers were first “originated in the United States, but this is not true.” 3 Nonetheless, the United States of America, Britain, and Germany all happened to create the computer during the same era. Germany was actually the first country that developed the computer. And the man that represented Germany is Konrad Zuse. Zuse “had the world’s first general-purpose programmable computer up and running” by the end of the year 1941. 3 Next came Bletchley Park, who represented Britain; he was a "top secret wartime establishment which was devoted to breaking the Wehrmacht's codes." 3 With the help of mathematicians and engineers, Park was able to build "an electronic computer for deciphering coded messages" called the Colossus.
3 By December 1943, "the first Colossus was installed" and put to use.3 Two years later, the United States came up with another system--the ENIAC. The ENIAC, also known as the Electronic Numerical Integrator and Computer, was built in University of Pennsylvania by John W. Mauchly and J. Presper Eckert in Novemeber of 1945. They continued building other machines to perfect the previous machines that had been developed.4 From this, the United States has become "the computer industry." 5 When one of their colleagues, John von Neuman who was known as an "unearthly genius", saw their inventions, Neuman began making huge contribution to computer design.5 "He enunciated the fundamental architectural principles to which subsequent generation of computers have adhered. For this reason, standard modern computers are known generically as von Neumann machines." 6 A few years later, in 1956, Neuman developed one of the most influential computer called the JOHNNIAC. The JOHNNIAC was an "early [effort] at AI programming." 6 This program was known as the Logic Theorists. Logic was used often in AI research.
"The ability to reason logicially is, obviously, a fundamental component of human intelligence; if computers are to achieve the status of artificial intelligences,
they too must be given the ability to search logically for a problem's solution." 6
From this point on, "the road to AI was rocky." The subject on "artifical intelligence was given its named by John McCarthy."9 McCarthy, one of the experts in computer, organized a conference called The Dartmouth Summer Research Project of Artificial Intelligence in 1956 dealing with artificial intelligence. Although the conference did not go as McCarthy hoped for, the studies in AI did not cease. As a matter of fact, the field in artificial intelligence became known to many universities: Carnegie Mellon, Stanford, MIT, and Edinburgh.9 Research on AI has continued ever since.
What is Artificial Intelligence
We have discussed the history behind artificial intelligence. But the question now is exactly is artificial intelligence? There are so many interpretation of what AI is and is not. Because of this, there is no "precise definition." 10 And "probably there will never be a clear [definition for artificial intelligence]." 11 The reason for not having a clear definition for AI is because those who are knowledgeable in AI are always trying to decide "where the boundary between AI and non-AI lies;" thus, they "do not always spell out their criteria" about artificial intelligence. 12
So what are the some of the definitions for artificial intelligence? The first one is defined by Boden. He defines artificial intelligence "as the study of how to build and/or program computers to enable them to do the sort of things that minds can do." 13 Another similar definition is defined by Marvin Minsky in 1968. He defines AI as "the science of making machines do things that would required intelligence if done by men." 14 This quote is widely known in the field of AI. This definition was then specified even more by Weizenbaum in 1976. Weizenbaum argued that "the overriding goal of AI is to create an artificial system which will equal or surpass a human being intelligence."15 A year later, Winston and Boden pointed out that artificial intelligence is "the use of computer programs and programming techniques to cast light on the principle of intelligence in general and human thoughts in particular."15 And in 1981, Haugeland had a different interpretation of AI. He simply believes that "intelligent beings are semantic engines -- in other words, automatic formal systems with interpretations under which they consistently make sense."15 To him and others like him, AI systems are merely "tools to perform specific tasks."15 However, if artificial intelligent systems have the minds to know when and how to "perform specific tasks," then they have the minds to "diagnose, advise, infer, and understand." 13 Thus, artificial intelligence systems cannot simply be just "tools." One can see already, the contrasting intereptations of AI. Therefore, there is no straightforward definition. One can literally go on trying to define this term.
AI supporters argue [. . .] that artificial intelligence is genuine intelligence: that human intelligence is merely the capacity of a naturally evolved organism to perform various sorts of operations--operations which are no different from those which might be performed by an artificial system.
On the other hand, intelligence is usually taken to be a feature of mentality: to be intelligent is to have a mind. To have a mind is to process a lot more properties than merely those associated with the capacity for intelligent performances. Just what is involved in having a mind may differ from species to species. [. . .] In the case of humans, having a mind seems to involve [. . .] being able to have sensuous awareness of objects and of one's own experience and physical states; being able to experience pain and pleasure, love, hate, fear, anger, ecstasy, serenity; being able to be creative, inspired, nauseated, ashamed, bored; being able to tell jokes and find them funny, to play, to be aroused, to be satisfied, to suffer.
How does it Work
So how does artificial intelligent systems work? First we must realize that the brain (the part of the intelligent system where the mind functions) often time does not carry out an action or perform a task immediately after the command is given; instead, the brain of the AI needs to do some processing. When AI is first created, "the memory is empty of concepts and response rules." The artificial intelligent system is built with a main objective. With this main objective, the AI is programmed to be able to come up for itself some sub-objectives, which are usually temporary objectives. The memory begins to fill up through its experiences. With its brain, the AI receives information from the senses (the parts of the intelligent system where they receive information from their environment--information outside of the system). The brain of the system then examine the information given to see if there are any concepts. If not, the system stores these information as elementary concepts. As the system receives more and more "information, expressed as concepts, the system builds up the present situation." Then the brain checks the memory to see if there are any response rules ("a field of storage that includes the present situation to which the rule is applicable [. . .] corresponding action") that may be related to the information. 17
For example, when the word students comes up, one would probably connect that word to the word teachers and/or schools. So if the senses of the system reports to the brain the word "students," the brain would check its memory to see if anything in the past would relate to the word students. Once the brain has found something that may relate to students, the system concludes that the concept of students would have a link concept to schools and/or teachers.
However, if the brain cannot find a link to the information it has received from the sense, it would form an elementary concept that would relate to the environment at that specific moment. After several encounters of elementary concepts, the brain rechecks to see if there are any response rule that may have any connection. The system does this by recalling similar experiences. With each experiences, the systems "records the present situation and the action that followed as a response rule." It later selects one of the best response rules and sends the actions to the limbs where the limbs carry out the action.17
After a long day of processing all these concepts, the artificial intelligent systems become inactive--the time they sleep. Like we, human beings, artificial intelligent systems need sleep in order to gather up all the information they have learn that day. During its inactivity, the system reviews the "response rules stored in its memory and perform some generalizations." Once they have made some generalizations, they have made response rules applicable to more situations.17 As artificial intelligent systems come across more and more different experiences, they will eventually learn how to make their own decisions.
Earlier on we have discussed the history behind artificial intelligence, defined what an artificial intelligence is, and explain how it works. Analyzing the ethics of artificial intelligence would be somewhat difficult without these information. To refresh our memory, the technological field has been unbelievable. It has now given us the opportunity to expand on the subject of artificial intelligence. The artificial intelligence system is a system that "learns how to act in order to reach its own objectives, that senses its environment, and using its experiences, selects an action and does it."17 The question now is whether or not AI is ethical.
To first analyze the ethics of any subject matter, one must determine what issues may affect the decision(s) that will be made. In the case of artificial intelligence, the first thing that usually comes to mind is that AIs are convieneint for mankind. It would make life much easier having the artificial intelligent systems to do people's works. Having artificial intelligence around would not only cut the tasks in half, but also it would save lots of valuable time that could be used to improve other areas in technology. Thus, this allows more opportunity to expand on one's knowledge that may benefit society.
However, with AIs becoming more common to society, people may become overly dependent on the artificial intelligence systems. This can be extremely harmful for society. For example, instead of taking the opportunity to expand on their knowledge and work productively, some people may just take this for granted. They would formulate the idea that because artificial intelligent systems are here just to ease the job, they could be lazy and become "coach potatoes." Not doing anything would be a waste in human intelligent.
In addition to this, depending on artificial intelligence systems can be dangerous if the systems were to be taken away from people. For example, we have become dependent on electricity in many areas of our lives. Now imagine what happens when the power goes out. Chaos! We would find it difficult to live without electricity. However, one may say that people back in the days have survived without electricity, so why can't we? The thing was that they were not depending on electrical appliants to do their tasks, and so they do not know of anything of electricity. Now the situation is turned around; once we have been dependent on electricity, we do not know of anything but electricity. Therefore, once we become dependent on AI, there is that risk of chaos if it were to be taken away from us.
Should the benefits of AI be taken away from those hardworkers just because of some foolish people?
Another thing that may affect the decision(s) that will be made is what would happen to the economy if AI was allowed. One thing that may possibly happen is that people may lose their jobs as the results of AI. The reason being is that artificial intelligence would be taking over people's old jobs, thus running out of things people can do. However, this should not be the case. There are always room for improvements; therefore, people can come up with other jobs to do that may help society even more. Another thing that would happen is that those-researchers, scientists, and inventors- who were involved in the field of artificial intelligence would be getting more money. Thus having more money resources that would allow them to research in other fields. This too can benefit society.
If people take complete advantage of what they got, good things can come from it--goods that may even benefit society as a whole.
Moreover, the field in technology would continue to increase exponentially. If further research on artificial intelligence were to be banned, the U.S. may become further behind in technology compared to the rest of the countries. Not only would the we be behind in technological fields, but also we would be taking away the rights of scientists, reserachers, and creators of artificial intelligence. These people have invested so much time and energy in this field, and to take it away this from them would be wrong. How would it be like if you were working on some projects for a very long time, and someone was to stop you from completing the project that you have started?
However banning artificial intelligence would prevent any further questions on where to draw the boundary line of what is and is not ethical. In banning any further research in artificial intelligence would mean to play safe of not crossing that boundary.
Is it better to play safe or take risk yet possibly benefiting from it?
One must also consider that once further researches is allowed in artificial intelligence, do artificial intelligent systems get their rights? What kind of rights would they get, if any? And at what types of AI will get their rights? One may say, AI should get their rights if they were made like human beings. But what makes an artificial intelligent systems human enough to gain their rights? When they have emotions? At the rate technology is going, it would not be surprising if AI do have emotions. So once they have emotions would they still gain the same rights as humans? Certainly AI did not come into the world to be destroyed. The main purpose (objective) of any intelligent system, which includes artficial intelligent systems, animals, and humans, is to survive. So how should AI be treated once they are here? There are many difficult ethical questions that need to be answer if further research is allowed.
One must also keep in mind what would happen if artificial intelligence were to surpass humans' intelligence? What would happen to mankind? Nobody would really want to create a system that would be better then himself. (Though this has happened in 1991. Arthur, the chessmaster, lost a chessgame to an AI system.) If the intelligence of artificial intelligent systems were to surpass the intelligence of humans, AI may have complete control over human beings. However, this is not the plan. In the Book of Genesis, God has given men the authority and responsibility of all things on earth.
Rule over the fish of the sea and the birds of the air
and over every living creature that moves on the ground.
Artificial intelligence systems would be considered the "living creature that moves on the ground." So if AI's intelligences surpass humans', how would people be able to fulfill this responsiblity? It would be disappointing to not be able to fulfill this responsibility just because one was "playing God." Men should not tamper with God's work for men do not have the right to create and destroy life.
From what has been said so far, we know that human beings now and in the future are major stakeholders. Another stakeholders are the scientists, reserachers, and inventors of AI. Last but not least, the artificial intelligence systems themselves are at stake.
In addition, we can see the different possible actions that could be carried out. The first and obvious possible action is to allow further research in the field of artificial intelligence without any restrictions. The second option is to completely ban any further studies on artificial intelligence. And the third options is to put restrictions in this field of study. We have three possible actions, the first two were discussed in the previous paragraph, so which do we choose? (Note there has not yet been any legal/policy issue in this area in general. The policy would vary amoung each organization.)
Although the first possible action--allowing further research without restrictions--may help improve technology, there are more harms than benefits like discussed in the paragraphs above. In addition, we run into the risk of crossing the boundary line in moral standards. However it granted the scientists, researchers, and creators of the AI field to continue with what they are doing. This action stresses more on the individual rights and fairness perspective rather than the utilitarian, common goods, and virtue perspectives.
On the other hand, the second possible action--ban completely--has no stress on rights and fairness perspectives. It simply just would not allow the researchers to continue researching. However, this action would stress on the virtue and common good perspectives. This action would develop moral virtues for society. Although it may be virtuous, society may lose some of the good benefits that may come out of it.
Choosing one of the two extreme possible actions would not be practical in society. The first one can be dangerous if go too far, and the second one violates people's rights. Therefore, the best decision that would benefit society is to compromise both of these: something in between the two that are not too extreme. Scientists, researchers, and inventors should be allowed to continue on with this field with some restrictions in order to allow them to expand on their knowledge at same time society would remain more virtuous than if they had no restrictions.
The technological advancements are incredible; one should take advantage of this and not take it for granted. And the best decision that would reflect this is to allow scientists, researchers, and creators of artificial intelligence to continue with what they have been doing. However, there needs to be some restrictions with a clear boundary line that would benefit everybody in society. The best decisions are the decisions that are best for everybody.
1. Otto Neumaier, Artficial Intelligence--The Case Against ed. Rainer Born, (London:
Routledge, 1989), 132.
2. Aaron Cook, Abhijit Khanna, and Maureen McDermott, "A Brief History of Artificial
Intelligence" Artificial Intelligence,
<http://bsc.edu/~spitts/cognitive/projects/briefhis.htm> (27 Feb. 2003).
3. Jack Copeland, Artificial Intelligence--A Philosophical Introduction (Cambridege:
TJ Press, 1998), 4.
4. Copeland, 5.
5. Copeland, 6.
6. Copeland, 7.
7. William R. Arnold and John S. Bowiee, Artificial Intelligence--A Personal,
Commonsense Journey (New York: Roxby Science Limited, 1987), 15.
8. Brief Timeline of Artificial Intelligence, <http://www-
scf.usc.edu/~burzlaff/project/Timeline.htm> (27 Feb. 2003).
9. Copeland, 9.
10. Arnold, 17.
11. J.A. Campbell, Artificial Intelligence--Human Effects ed. M. Yazdani and A.
Narayanan, (New York: Halsted Press, 1985), 249.
12. Campbell, 270.
13. Margaret A. Boden, The Philosophy of Artificial Intelligence (Oxford: Oxford
University Press, 1992), 1.
14. Tom Stonier, Beyond Information—The Natural History of Intelligence (London:
Springer-Verlag, 1992), 107.
15. Neumaier, 135.
16. Steve Torrance, "Ethics, Mind and Artifice," Artificial Intelligence for Society.
ed. Karamjit S. Gill, (Chichester: John Wiley & Sons, 1986), 56.
17. Walter Fritz, Intelligent Systems and Their Societies.
<http://www.anice.net.ar/intsyst/index.htm> (27 Feb. 2003).
Aleksander, Igor and Burnett, Piers. Thinking Machines—The Search for Artificial Intelligence. New York: Roxby Science Limited, 1987.
Arnold, R. William and Bowie, S. John. Artificial Intelligence—A Personal, Commonsense Journey. Englewood Cliffs: Pentice-Hall, Inc., 1986
Boden, Margaret A. The Philosophy of Artificial Intelligence. Oxford: Oxford University Press, 1992.
Born, Rainer. Artificial Intelligence—The Case Against. London: Routledge, 1989.
Collins, H.M. Artificial Expert—Social Knowledge and Intelligent Machines. Massachusetts: The MIT Press, 1990.
Cook, Aaron, Abhijit Khanna, and Maureen McDermott. Artificial Intelligence. 27 Feb. 2003. http://bsc.edu/%7Espitts/cognitive/projects/index.htm.
Copeland, Jack. Artificial Intelligent—A Philosophical Introduction. Cambridge: TJ Press, 1998.
Crevier, Daniel. AI—The Tumultuous History of The Search for Artificial Intelligence. New York: Harper Collins Publishers, 1993.
Fritz, Walter. Intelligent Systems and Their Societies. 27 Feb. 2003. http://www.anice.net.ar/intsyst/index.htm.
Gill, Karamjit S. Artificial Intelligence for Society. Chichester: John Wiley & Sons, 1986.
Stonier, Tom. Beyond Information—The Natural History of Intelligence. London: Springer-Verlag, 1992.
Yazdani, M. and A. Narayanan. Artificial Intelligence—Human Effects. New York: Halsted Press, 1985.