Arguments for Artificial Intelligence

De La Salle University

Arguments on Artificial Intelligence

Niccolo Marcus T. Defiesta
INTFILO V24
Dr. Noelle Leslie
Submitted on August 27, 2014
What is Artificial Intelligence (AI)? By definition, AI is the intelligence exhibited or exercise of though by a machine or software. Philosophically, the most important question is “Can a machine really think”? This question goes beyond the realms of science because there are religious, legal, and ethical implications with the concept of AI. We humans are rational beings; beings who place our morals in high regards, such as the conscience to have courage or to not cheat during a test. We humans are also sentient beings that are subject to pain, suffering, love, and hate. So can a machine really “think” in regards to us humans? Since machines such as a simple calculator to a laptop have the appearance of intelligence, the question arises: “Are they really thinking” or are they just mindless machines set only to the programs they are designed to do? If they are thinking, then should they have the same rights and treatments as us humans? What are the political, ethical, and religious changes that may arise? The movie, The Matrix, is a movie we tackled regarding AI, but that’s just fiction – or is it?

In this paper, I will discuss the arguments for AI and the counter-arguments against AI and provide facts and information for both stands. I will also discuss the political, religious, ethical, and moral implication of this issue. I will first evaluate and explain what exactly merits an entity as intelligent and where the “felt qualities” belong and how they are important to the title of intelligence. I then explain the famous Turing Test and how it is used to test a machines capability to exhibit intelligent behavior equivalent to, or indistinguishable from that of a human. Afterwards, I will give a brief history and summary of the origins of AI, as well as the arguments for AI and the counter-arguments against AI. Finally, I will deduce all the information collated and give my own personal and subjective answer whether AI is truly possible.

1. Intelligence
a. What is Intelligence?
Intelligence can be defined as “The ability to deal with cognitive complexity” (Linda Gottfredson). Qualities such as the ability to construct correct reasoning, to plan ahead, to solve problems quickly, and to learn and evolve from experience (etc.) are all important indicators of intelligence. Thus, the better the ability an entity is at these cognitive abilities, the higher the intelligence it is said to be.

Intelligence among machines seems to be shown everywhere around us. From chess programs to self-cleaning machines, it seems as though they have some low-level intelligence inside of them; or do they? The majority of machines that are used today, are programmed to specific programs or instructions that are embedded in their programming. Moore’s law (or observation) states that the number of transistors in a circuit roughly doubles every two years; this prediction has been proven to be accurate, meaning that the growth and development of the technology sector is growing at an exponential rate – just look at the first computer (German Z3) and compare it to the world’s leading computer system (Tianhe-2).

2. The Famous Turing Test: Am I A Machine?
The Turing Test was created by Alan Turing in his 1950 paper – Computing Machinery and Intelligence. In this test, Turing suggested that if a computer can talk like a human, then it is intelligent. Turing classified the test as an “imitation game” whose original version was played by three people, (1) a Man, (2) a Woman, and (3) an Interrogator who could also be played by either sex. The test is setup wherein the interrogator stays in a room apart from the Man and the Woman. During the test, the interrogator will ask questions about anything to both the Man and the Woman. The goal of the interrogator is to determine who the Man is, while the Woman tries to trick the Interrogator into mistaking her as the Man.

Turing then tweaks the mechanics of the game by changing the Woman to a Machine. Turing then asks the question, “What will happen when a machine takes the part of A in this game?” This then asks the question if the interrogator will decide wrongly as much as when it was played by humans only.
Illustration of the test: A B

A – Where the test is done by humans only. (1) The Man tries to help the Interrogator in making the correct decision. (2) The Woman tries to trick the Interrogator in making the wrong decisions. (3) The Interrogator asks both the Man and the Woman questions, then decides which one the Man is.

B – Where (2) is replaced by a Machine. (1) The Man tries to help the Interrogator in making the correct decision. (2) The Machine tries to trick the Interrogator in making the wrong decisions. (3) The Interrogator asks both the Man and the Machine questions, then decides which one the Man is.

Turing proposed this test as a ‘qualifying test’ for human-level intelligence. Though this test is used for testing for human-level intelligence, it is not a disqualifying test for intelligence. Many A.I. advocates state that the Turing Tests rubrics for classifying whether a machine is intelligent or not in this test is lacking, it still serves as a base-plate for classifying and understanding a line for which intelligence can be judged. But there are still advocates to this test that stay true to the claim that if a machine were to pass this test, then they could be classified as intelligent.

3. AI Today
I. A Little History
We owe the invention of computers to none other than the 19th century English mathematics professor named Charles Babbage. Babbage is accredited to being the designer of the Analytical Engine which is the design that the basic framework of the computers of today are based on.

The genealogy of the computers can be divided into three generations. Each generation developed a newer and better computer or made game changing improvements to the existing computer.

First Generation: 1937 – 1946 – In 1937, the first electronic digital computer was created by Dr. John V. Atanasoff and Clifford Bay. The machine was called the Atanasoff-Berry Computer, or ABC. In 1943, the military developed their own electronic computer named Colossus. And in 1946, the first-general purpose computer was built, the Electronic Numerical Integrator and Computer or ENIAC was built. The ENIAC was said to have weighed over 30 tones, and had over 18,000 vacuum tubes for processing. Despite such a large machine, the ENIAC had no operating system and could only perform single tasks.

Second Generation: 1947 – 1962 – The reason for the change in generation can be attributed to the use of transistors instead of vacuum tubes. During this period, the first computer that was readily available for the public was also issued, the Universal Automatic Computer, or UNIVAC1. As well as the development of the IBM 600 and 700 series was released which is iconic as one of the first true computers to be used in the business world. The use of disks and tapes were also prominent during this time, as well as the development of over 100 computer programming languages.

Third Generation: 1963 – Present – The year 1963 started a computer revolution with the creation of the integrated circuit. With the integrated circuit, computers became smaller and smaller and wiped out the use of humungous and clunky machines, leading to the development of computers that could be used for home purposes. In 1980, MS-Dos was built and in 1981 IBM created the first personal computer, or PC for home and office uses for individuals. In 1984, the Macintosh computer was born which revolutionized computers with an icon driven interface. And in the 90’s, Microsoft gave birth to the Windows operating system.

II. Practicality
The faster the operational capacity of the machine the greater the computational power the machine has. From cogwheels to circuit boards, improvements in computer performance has skyrocketed, unleashing a technological boom to development of faster, more powerful machines.

This boom in technology can be attributed to Moore’s law which was described in the early pages of this paper. To refresh your memory, Moore’s law states that the density of circuits roughly doubles every two years. But this observation is only applicable to the hardware of a machine, what about the software? The development of better and more powerful software is far behind the progress of hardware. But even with the lagging progress of software, we still have managed to create programs that could be deemed intelligent, if done by humans or animals. Machines are now capable to do many things formerly done by animals that could be considered as intelligent, like, seeking for objects or anomalies in the environment, detecting harmful substances or bacteria, and tracking individuals or hazards; seeming evidence of basic-level AI. Machines are capable of doing human activities that could be considered intelligent, such as, playing games, planning, performing complex math problems, and in a way learning from ‘experience’; seeming evidence of human-level AI.

b. Arguments for AI
I. Low-Level AI
Some machines today could be described as having mental abilities. An example of such ability is a thermostat. Room thermostats change the temperature of a room by monitoring the temperature and adjusting it to the right temperature. But if you were to place an ice cube in front of the thermos, then the machine would react to this by lowering the temperature, despite the room not actually being as cold as the ice that was placed in front of it. Monitoring, adapting, and changing things seem to be indicative of mental processes or conditions, marked by their intentionality. Humans naturally have a higher level of intelligence, but have as well low-level intelligence just like how some animals and machines do. We like to characterize machines with low-level intelligence: we say that they can connect automatically for WiFi spots, they try to connect with other devices, and detect key presses in a computer or cellphone, etc. But what about a higher level intelligence like how the Turing Test and classical view of AI conceive it to be? The Turing Test and classical view of AI are more concerned with higher levels of intelligence, such as the following examples:
Mathematics Neural Networks
Robots
Strategic Planning
Games
Networks Mathematics
Speech Conversion

Despite having an abundance of low-level intelligence machines and some higher-level intelligence machines, we are still far from reaching the characters of C3PO and the likes. In particular, the challenge Alan Turing proposed still remains unmet. Whether or not we will ever reach that point, is unknown.

4. Arguments Against AI
a. Consciousness
... people always tell me it was very hard to define consciousness, but I think if you 're just looking for the kind of commonsense definition that you get at the beginning of the investigation, and not at the hard-nosed scientific definition that comes at the end, it 's not hard to give a commonsense definition of consciousness. Consciousness consists of those states of feeling or sentience or awareness; they 're the kind of things that start in the morning when you wake up from a dreamless sleep; they go on all day until you get knocked on the head and become unconscious, or you fall asleep again, or you die, or otherwise your consciousness is shut down. And on this account, I hope it 's obvious, dreams are a form of consciousness. Now, that is the definition of consciousness.
By Alan Saunders, May 20 2006

This excerpt from Alan Saunders is the main argument from anti-AI advocates on the subject of Consciousness. What Saunders was trying to say here, is that machines lack the sense of felt qualia. If you were to input a program and receptors to detect environmental changes in a room, this would still not be adequate in supplying them the sensations of heat, cold, dry, or wet conditions. The argument is, is that these sensations are what consciousness is made up to be.

b. The Chinese Room Argument

The Chinese room argument is a thought experiment of John Searle (1980a) and associated (1984) derivation. It is one of the best known and widely credited counters to claims of artificial intelligence (AI)—that is, to claims that computers do or at least can (someday might) think. According to Searle’s original presentation, the argument is based on two key claims: brains cause minds and syntax doesn’t suffice for semantics. Its target is what Searle dubs “strong AI.” According to strong AI, Searle says, “the computer is not merely a tool in the study of the mind, rather the appropriately programmed computer really is a mind in the sense that computers given the right programs can be literally said to understand and have other cognitive states” (1980a, p. 417). Searle contrasts strong AI with “weak AI.” According to weak AI, computers just simulate thought, their seeming understanding isn’t real understanding (just as-if), their seeming calculation is only as-if calculation, etc. Nevertheless, computer simulation is useful for studying the mind (as for studying the weather and other things).

By Larry Hauser, 2005

This excerpt from Larry Hauser on the CRE simply states that – even if a computer speaks and communicates effectively in Chinese, it does not understand Chinese, rather, it only simulates that knowledge.

c. Originality
The Analytical Engine has no pretensions whatever to originate anything. It can do whatever we know how to order it to perform. It can follow analysis; but it has no power of anticipating any analytical relations or truths.
By Lady Lovelace, 1842

This excerpt from Lady Lovelace is one of most famous objections against AI. What Lovelace here is saying is that – machines have neither free will or originality. That machines can only do whatever information or rules that it was given, and that we as humans are autonomous, and machines are automata in nature.

5. Conclusion
While currently there is still no concrete definition of what intelligence is, scientists and philosophers are slowly, but continually, coming onto an agreement to what intelligence really is. But with the ever expanding field of technology, the development of low-level AI will inevitably bring about the growth of higher-level AI machines. Despite the boom in technology, arguments for behavioral evidence against AI still currently overshadows the scientific evidence for AI, but as more research is poured into the development of AI, the scientific evidence will inevitably overcome the behavioral evidence against it – if at all the day comes. At the current level of AI (low-level), machines are able to do processes and functions that would label them as intelligent if done by humans. But as we approach what Alan Turing has proposed, or at least the qualities for basic human level intelligence, if at all it comes, then it should entitle machines to be deemed intelligent, from what we currently know.

6. References
Dreyfus, Hubert. 1998. “Intelligence without Representation.”
Fodor, Jerry A. 1975. The Language of Thought. New York: Thomas Y. Crowell.
Lenat, D. B., M. Prakash, and M. Shepherd. 1986. Cyc: using common sense knowledge to overcome brittleness and knowledge acquisition bottlenecks. AI Magazine, 6(4).
McCarthy, John. 1979. “Ascribing Mental Qualities to Machines.” In Ringle, M. (ed.), Philosophical Perspectives in Artificial Intelligence. Harvester Press.
Moore, G. 1965. “Cramming More Components onto Integrated Circuits.” Electronics 38: 8.
Searle, J. R. 1980. “Minds, Brains, and Programs.” Behavioral and Brain Sciences 3
Turing, Alan M. 1950. Computing machinery and intelligence

References: Dreyfus, Hubert. 1998. “Intelligence without Representation.” Fodor, Jerry A. 1975. The Language of Thought. New York: Thomas Y. Crowell. Lenat, D. B., M. Prakash, and M. Shepherd. 1986. Cyc: using common sense knowledge to overcome brittleness and knowledge acquisition bottlenecks. AI Magazine, 6(4). McCarthy, John. 1979. “Ascribing Mental Qualities to Machines.” In Ringle, M. (ed.), Philosophical Perspectives in Artificial Intelligence. Harvester Press. Moore, G. 1965. “Cramming More Components onto Integrated Circuits.” Electronics 38: 8. Searle, J. R. 1980. “Minds, Brains, and Programs.” Behavioral and Brain Sciences 3 Turing, Alan M. 1950. Computing machinery and intelligence