artificial intelligence overview

42
TABLE OF CONTENTS Introduction 1 Historical Development 1 Main Issues of AI 3 Goals of Artificial Intelligence 3 Thinking Machines 6 Regulation and Ethics 8 Stakeholders 9 Government 9 Businesses 10 Society 11 Biblical Approach 12 Bibliography 16 Appendix 20 End Notes24

Upload: independent

Post on 01-Apr-2023

1 views

Category:

Documents


0 download

TRANSCRIPT

TABLE OF CONTENTS

Introduction 1

Historical Development 1

Main Issues of AI 3

Goals of Artificial Intelligence 3

Thinking Machines 6

Regulation and Ethics 8

Stakeholders 9

Government 9

Businesses 10

Society 11

Biblical Approach 12

Bibliography 16

Appendix 20

End Notes24

1

Understanding Artificial Intelligence

Introduction

On November 5, 2014, To-Robo, an artificial intelligence

(AI) agent outperformed the average of high school students on a

standardized test. To-Robo surpassed most students on the English

section of the exam, scoring a 95 out of 200 points compared to

the average of 93.1 for Japanese students.1 Even though To-Robo

struggles in other subjects such as object recognition and

politics it might one day be able to outperform humans in most

areas.

What is intelligence? Are machines capable of intelligent

behavior? Should AI focus on human biology to create intelligent

agents, or is AI unrelated to human intelligence?2 These

questions are the basis for the field that is known as AI today.

The purpose of this research is to provide a holistic approach to

2

the understanding of the field of AI. The following terms defined

are important for the understanding of AI:

Artificial Intelligence: Is the field of science that aims to create machines that possess intelligent behavior.

Algorithm: An algorithm is a step by step procedure or instructions given to accomplish a task.3 In the case of AI an algorithm will most likely be an equation or set of equations coded into a machine to accomplish a certain goal.

AI Agent: Stands for any new technology or procedure that isgenerated by the AI field.

Strong AI (General AI): For some experts, strong AI is the ultimate goal of the field which aims to produce an AI agentpossessing the same capabilities and intelligence a human does.4

Weak AI: In contrast with Strong AI, weak AI is any technology developed that is capable of accomplishing intelligent tasks but lacks the overarching purpose of general intelligence.

Moore’s Law: Theory proposed by cofounder of Intel, Gordon E. Moore. The theory states, processing power and speed of acircuit will double every eighteen months.5 This trend noticed in 1958 has held true to this day.

Historical Development

Humanity has always been fascinated by the possibility of

creating intelligent beings. Before AI emerged as a field,

writers, philosophers, and scientists were already speculating on

the possibility of intelligent machines. Mary Shelly is famous

for her book Frankenstein; published in 1818.6 The books main

event about a scientist who was able to create artificial life.

3

Isaac Asimov, a professor of biochemistry and fiction writer, is

perhaps most famous for his book series called I Robot, published

in 1950. Asimov is recognized for conceptualizing the possibility

of general intelligence, and creating a basic framework of rules

to govern all artificial agents.7 (See Appendix A, Asimov’s three

Laws of Robotics).

The possibility of an artificial agent had been discussed

years prior to the rise of the field. Alan Turing was one of the

first to introduce the possibility of intelligent machines.

Turing published his study called Computing Machinery and Intelligence

in 1950. In this book Turing poses the question: Can machines

imitate intelligent thinking?8 Turing, also proposed the Turing

Test as a means of assessing the intelligence of future AI

agents. The test is famous for devising a method to help prove if

a machine could be considered intelligent or not. Turing proposed

that through advances in technology it might be possible to

produce a machine that is capable of providing response and

solutions that accurately imitate the answers a human being would

give.9 Thus, this ability would constitute the machine to be

4

intelligent in the sense that it is capable of producing coherent

answers indistinguishable from human responses.

However, it was not until 1956 during a conference in

Dartmouth College that John McCarthy, computer scientist and

researcher, proposed the term Artificial Intelligence for the

first time in history.10 In this conference the term AI was used

for the first time along with the presentation of one of the

first AI programs. Herbert Simon, and Allen Newell presented the

Logic Theorist (LT) in 1956. The LT was a program that searched

for answers based on trial and error. McCarthy, was also the

first person to create a programing language for AI in 1957

called LISP.11 LISP made it easier to find a common language to

make new programs and generated several advances in the field.

There have also been setbacks for AI including a period in

the 1980s known as the AI winter. This period is known for an

overly optimistic view of AI and its capabilities. The Defense

Advanced Research Projects Agency (DARPA) also began to cut

funding for projects due to lack of results.12 During the 1990s

there was a resurgence of AI. For the first time, in 1997, a

computerized chess program (Deep Blue) beat the reigning world

5

champion Garry Kasparov.13 In the early 2000s robotic pets first

became available to purchase by the general public. From its

humble beginnings, AI has made tremendous contributions and

technological advances in the world. Understanding its history

allows for analysis of the implications that AI will represent

for society in the future.

Main Issues of AI

The goal of AI is to create machines that possess

intelligent behavior. This section is divided into two main

parts. The goals of the AI field will be discussed along with

strategies implemented by different scientists. The different

ideologies of whether or not it is possible to create thinking

machines will be reviewed along with two main problems of

creating a truly intelligent AI agent.

Goals of Artificial Intelligence

To create a truly strong AI agent researchers identified the

main areas that need to be developed. A truly intelligent agent

should have the ability to reason like humans do. AI should also

have vast knowledge on extensive topics. An AI agent should be

able to plan, set goals, and have a way to accomplish these

6

goals. An intelligent being should have the ability to learn new

information through experience in the real world, and be able to

communicate solutions verbally. Researchers also discuss the fact

that an AI agent has to have mobility (robotics), and be able to

manipulate objects, and have visual abilities just like humans

do.14 To accomplish these goals AI research has taken several

different paths. The most prominent branches of AI are discussed

in the following sections.

Symbolic AI

The possibility of Strong AI stems from the idea of

functionalism. Functionalism interprets the human mind and

conscious thoughts as a series of steps and symbols.15 Scientists

believe machines could also be able to produce intelligent

behavior similar humans through the use of algorithms and

heuristics. A notable success for symbolic AI was the development

of the system ELIZA by AI researcher Joseph Weizenbaum in 1965.16

Designed to hold conversations using a text interface, ELIZA is

scripted using a symbolic framework that looks for patterns in

the users inputs (See Appendix B, Conversation with ELIZA).17 The

development of ELIZA enabled to creation of knowledge-based

7

systems such as Deep Blue, the chess playing machine discussed

earlier capable of beating grand master level players.

A Knowledge-Based (also known as an expert system) system is

a specialized symbolic AI that is constructed by information

provided by experts on a specific field like medicine or

accounting. Knowledge in the form of facts is stored in a

database, which can usually be accessed through questions

postured by a user of the system (See Appendix C, Structure of an

Expert System). One of the earliest examples of an expert system

was developed by researcher Ted Shortlife in 1974. This AI agent

was called MYCIN.18 MYCIN was composed to be and aid in giving

medical diagnosis and suggest medical treatments. MYCIN proved to

be an extremely efficient tool. For example, “In one formal test,

a judging panel rated MYCIN’s prescription as correct 65% of the

time compared to 42.5-62.5% for human specialists presented with

the same cases.” 19

Just like MYCIN there are also powerful expert systems being

developed today. Apple Inc.’s Siri is an example of an everyday

expert system used by people on their phones. IBM also developed

a system called Watson in 2011. Watson was able to win a Jeopardy

8

competition, not by understanding what was asked but with raw

computational power.20 Even though expert systems are powerful

they lack conscious behavior and have no real understanding of

the answers that are given. An expert system might give a correct

answer, but it does not mean it understands the consequences of

the solution. This type of program is known as weak AI.

Neural Networks

The second approach to AI is the development of neural

networks, first conceptualized by Warren McCulloch and Walter

Pitts in 1943.21 Researchers of Neural Networks try to recreate

the activities of the human brain by the function neurons perform

in transferring information. The first application of a neural

network is called the perceptron, created by scientist Frank

Rosenblatt in 1950.22 The perceptron is an algorithmic function

created to follow the basic design of a real neuron. Neural

Networks require computers with great amount of raw processing

power which is why they fell out of favor with researchers in the

early years. Applications with neural networks surfaced as

computers gained more computational power in the 1980s.23

9

Genetic Algorithms (GA), is a branch of the neural network

approach that stems from evolutionary ideas. Scientists attempt

to create algorithms that copy evolutionary processes by process

of crossover and elimination.24 With this process GAs are able to

arrive to a best-fit solution by elimination, and arrive to an

optimum solution.

Robotics

In contrast to symbolic and neural approaches, robotics has

the ability to create AI agents that allow mobility. This

mobility also allows scientists to study how an intelligent agent

(robot) behaves when faced with uncertainties or unanticipated

situations that occur in a natural environment.25 Some scientists

say that it is not efficient to build robots because of the high

cost; instead computer programs should be able to model the same

abilities that robots would have in a real life situation.

Thinking Machines

When creating AI the following questions is considered: Are

machines capable of intelligent behavior? There are two differing

arguments supported by researchers. First, those who believe

10

general AI is possible and inevitable; this is based on the

theory of functionalism and Moore’s law. Second, the position

saying that achieving general AI is ultimately impossible.

Proponents of the possibility of general AI base their ideas

on Moore’s law, and the belief of a possible technological

singularity. General AI proponents believe that in the near

future machines will become so powerful (Moore’s law) that they

will become a super intelligent agent that is capable of

conscious human-like thought. This event will lead to a

singularity where machines will be so intelligent that they will

be able to reinvent themselves. One of the main scientists in

favor of the possibility of the singularity is Ray Kurzweil,

current director of the department of Artificial Intelligence for

Google Inc.26 Kurzweil popularized the idea of general AI, and of

the possibility of the singularity. If this were to happen,

Kurzweil also argues that human beings would be able to merge

with existing technologies forming a sort of cyborg mixture

between human and machine. What Kurzweil is advocating is the

possibility of being able to live on forever with the aid of

advanced AI; this view is also known as transhumanism. With this

11

view, humans will transition minds to advanced intelligent

systems.

The possibility of AI is also fueled by the idea that the

human mind processes information and ideas in terms of symbols

and ordered step-by-step procedures (Functionalism). Early

developments of AI through symbolic AI followed this train of

thought and proposed that through heuristics a machine could

eventually become intelligent. With enough processing power

(Moore’s Law), future machines could be able to learn by

themselves and gain some sort of consciousness.

Opponents of general AI disagree with Kurzweil’s ideas that

have become popular in recent years. John Searle, professor of

philosophy at the UC Berkeley, states that general AI is utterly

impossible. Searle expresses that computer programs are basically

algorithms that by analyzing inputs can be programmed to carry

out specific functions. This means that, the AI system does not

need to understand the underlying conscious thought of an action,

but rather be programmed to carry out a certain expression.27

This argument is famously known as the Chinese room argument. The

argument also states that, one cannot simply program a computer

12

to express consciousness like a human does.28 Thus, a true

general intelligent AI is not possible in the sense that the AI

agent will not understand the reason why it is performing certain

tasks.

Hubert Dreyfus, professor of philosophy at the UC Berkley,

states that there are two main problems with traditional or

symbolic AI approaches. These two issues are that computer

programs lack the understanding of common sense and relevance. In

the early 60s, AI researches embarked on trying to create

conscious thought on programs with algorithms and symbols. What

researches found is that in spite of programs being able to solve

problems and use algorithms to find new solutions they lacked

common sense. A program cannot understand why the solution

provided is the right one, it arrives to the solution through

mindless calculation. The issue of relevance (Frame Problem) is also

present. AI systems cannot differentiate which data is relevant

in a specific situation. Relevance as used by human beings is

easy to understand because humans analyze their surroundings and

are able to apply relevant facts to a specific situation. The

same is not true for computer systems.29 A computer systems with

13

millions of data would have trouble selecting a relevant frame of

data to apply to a specific situation without being told to do so

by a programmer. From these problems it is evident that all of

the AI agents that are present today are considered to be weak AI

agents. There is no evidence that strong AI is currently a

possibility with existing technologies, but rather the belief of

general AI is mere speculation with no hard facts to prove it.

Regulation and Ethics

Currently there are no existing governmental entities that

provide oversight for the field of Artificial Intelligence.

However, products arising from AI research are generally subject

to regulation. Government takes on a reactive role to provide

regulation as new AI products arise. For example, Google Inc. is

currently producing autonomous self-driving vehicles. Several

states that have already passed regulation for Google’s self-

driving vehicle, as well as other future automated automobiles

(See Appendix D, California Bill 1298 regarding autonomous

vehicles).

AI is a concern for regulators when it comes to privacy

laws, since AI agents pose an increasing danger when it comes to

14

safeguarding information. The US government is currently

considering threats to private and public entity security when it

comes to privacy and information sharing. On July 10, 2014 the US

Congress proposed the Cybersecurity Information Sharing Act

(CISA). CISA required private and public entities to propose

plans for information security in dealing with threats, and

vulnerabilities to technology.

If general AI is indeed a possibility, creating an ethical

framework for an AI agent is equally important to passing

regulation. Elon Musk, owner of Tesla Motors, argues that AI

could be one of the biggest threats to human existence if not

regulated properly.30 The branch of AI called Machine Ethics was

created to consider the ethical implications of general AI.

Machine Ethics classifies two types of AI, explicit and implicit

agents. An implicit intelligent agent would act according to the

behaviors programmed in the machine by the creator, assuming the

creator follows correct ethical standards. Explicit agents would be

able to make ethical decisions based on an agreed upon code of

ethics by the AI field.31 The desirable outcome of machine ethics

is to create an explicit AI agent, because the laws that govern

15

each agent would not depend on individual scientist beliefs but

rather on an agreed upon code of ethics.

Stakeholders

AI is everywhere, all technology new an old can be traced

back and related to an advancement in a branch of AI. Even though

general AI is not possible today, new technologies can have

tremendous benefits for society if handled with care. As a whole,

AI affects civilization in a global scale; consequently, the

following important stakeholders have been identified.

The government has always had a special interest in the

advancements of AI. Since the 1960s the government has provided

monetary support and facilities to develop new technologies in

the form of the DARPA organization. A main concern for the

government is to provide advanced technology for national

defense, and AI agents for assistance during war time. Some of

the arguments against AI dedicated to national defense is the fact

that advanced AI could be used to generate mass destruction.32

The threat itself would not be towards the AI agent but, towards

the world powers that would use these technologies to coerce

society and generate chaos. Bill Joy, chief technologist for sun

16

microsystems, argues that humans will not be able to control

future technologies which will cause confusion an increase

violence. Joy, advocates for disarmament and prohibition of

technologies that would lead to general AI, and autonomous

weapons.33

Supporters for autonomous weapons acknowledge the fact that

dangers exist when handling AI. Supporters suggest that, “Instead

of prohibiting or heavily regulating artificial intelligence, the

United States should support civilian research into a kind of AI

that will not endanger humans—a so-called friendly AI.”34 As

discussed with machine ethics, the creation of friendly AI would

largely depend on successful implementation of an ethical

framework for all future AI agents. A successful friendly AI

would be beneficial for humankind, especially when it comes to

national defense. The lives of human soldiers will not be

endangered with developments of technologies such as autonomous

drones.

Businesses have been a great beneficiary of artificial

intelligence. The benefits AI has given businesses are increased

efficiency in the workplace, enhanced security with work

17

procedures, and improved decision making.35 Advances in computing

has increased the efficiency of businesses to analyze data and

assess the position of a company. Technology has enabled

employees to perform better at their jobs, and increasing

profits. For example, the availability of computers for an office

enables employees to handle heavier workloads than when these

were not widely available for use. AI has eased data analysis

needs for businesses. Data mining is an example of an AI

technology that aids analysts with pattern, and trend recognition

of company data. New machinery has improved working conditions

for manufacturing companies, and eased the work employees have to

perform.

The challenge businesses face are related to the failure of

autonomous systems, and protection of information.36 If a machine

or a computer system fails, a company needs to have security

measures in place to correct a potential crisis and reduce

collateral damage. Businesses today also need to consider

cybersecurity to protect sensitive information from unauthorized

access. The biggest challenge for businesses is to stay

18

competitive with new technologies and being able to protect their

information.

Perhaps the most important stakeholder for the field of AI

is society, and the effects new technologies can have on the

community. There are differing views related to the benefits, and

repercussions that the generation of AI can have on society.

Effects can be seen in the areas of employment opportunities, job

displacement, and changing communication between cultures.

Opponents say that AI can be good for some social classes and bad

for others. For example as low skilled paying jobs like item

assembly or packaging become increasingly automated the industry

would experience a decrease in job openings. New jobs available

would displace the low-skilled workers (creating unemployment) in

favor of an educated work-force to maintain these machines. Even

though these displaced workers could find work somewhere else, it

is not a guarantee that the new job would generate the same

satisfaction and provide for the needs of these individuals.37

Opponents also discuss the fact that new technologies can

displace educated high-skilled workers. Paul Krugman, Nobel Prize

winner and professor of economics, expanded his idea that in

19

previous times unemployment caused by new technologies was solved

with intensive education of lower skilled workers. This method of

intensive education will not solve newer technological challenges

posed by AI.38 This is evident today, in the accounting field. In

the past a bookkeeper job was considered to be a technical job,

requiring education and experience. Today, newer accounting

systems, have simplified this job to an extent that a person with

little expertise can perform what used to be a high-skilled

position.

Scholars and economists who approve the creation of AI

contend that, “workers may be expelled from a company or a

sector, but sooner or later they will be hired by other companies

or reabsorbed by a different economic sector.”39 New technologies

will create temporary unemployment, but this change is never

irreversible since employment rates have always fluctuated. If

new technological advances increased structural unemployment by

now all the population would be unemployed because there has been

increasing technologies for over two centuries.40 There is always

a concern for increased unemployment, but AI is not the main

reason for this change. AI will restructure jobs and might create

20

temporary unemployment, but in general technology benefits

society in much greater ways.

Creation of new technologies, such as the invention of the

tractor have eased the work of farmers. Airplanes, and electronic

mail have made communication between cultures easier than ever

before. An important consideration is that increased connectivity

also decreases privacy. The complaints that users of technology

have is that businesses, government, and third party users can

have too much access to private information. Privacy will always

be an issue, especially with the nature of society for increased

communication and information sharing.

Biblical Approach

AI aims to create machines that possess intelligent

behavior. Some scientists such as Ray Kurzweil advocate for the

possibility of general artificial intelligence. Others, such as

Dreyfus, say general AI is impossible, and that the technology

produced by AI is considered to be weak AI. Nevertheless, both

scientists and the general public are sometimes fascinated by the

possibility of creating autonomous artificial agents

(specifically the possibility of the singularity). These AI

21

agents according to Kurzweil, will have the ability to reason and

behave like humans. We have seen the different approaches taken

by AI in creating an intelligent being. These approaches are the

symbolic approach, and the biological approach (Neural Networks).

Still after more than sixty years scientists have not been able

to create a machine that has common sense, and that is able to

reason like humans do. It is evident from the field of Machine

Ethics, that creating friendly AI is a priority, since an ethical

framework is essential for any new developments towards general

AI.

As a Christian I want to concentrate on the possibility of

creating a strong AI agent. Strong AI claims that it is working

towards producing a machine that is truly intelligent, human-

like, having a consciousness and being able to make ethical

decisions just like humans do. Machines might be able to imitate

intelligent behavior (computation, problem solving), but I do not

believe that it is possible to create AI with the ability to

reason and discern knowledge just like humans do. Referring to

intelligence in AI, I want to stress the fact that human

intelligence is completely different from machine intelligence.

22

In his publication, Computing Machinery and Intelligence, Alan Turing

acknowledges that his intention for AI was to create a machine

that is able to play the “imitation game”.41 To be considered

intelligent I will suggest that a machine does not need to, and

cannot, have conscious thought. An intelligent machine can imitate

intelligence with the use of databases and algorithms.

I believe that conscious thought and intelligence in humans

comes from God. Genesis 1:26 says, “And God said, Let us make man in our

image, after our likeness: and let them have dominion… and over every creeping thing

that creepeth upon the earth.” God made us in his image with a special

purpose in this earth. Clearly, God created us to be superior to

all of his creation, as everything created in the earth is given

for us by God to administer. Genesis 2:7 also states that God

“breathed into his nostrils the breath of life; and man became a living soul.” In the

Bible, the creation of human beings is the only time were it says

that God breathed life into his creation. By this verse we can

understand that our consciousness and ability to reason is not a

mere physical product, but a spiritual one. Thus, consciousness

if not provided by God is an impossible feat to accomplish. Job

33:4 shows that life comes from the Lord, “The spirit of God hath made

23

me, and the breath of the Almighty hath given me life.” God gave us both a

physical body, and a spirit which allows us to have discernment

of good and evil. Ultimately we are created for God’s glory, and

our ability to have conscious thought should be purposed to serve

the Lord (Thou art worthy, O Lord, to receive glory and honour and power: for thou

hast created all things, and for thy pleasure they are and were created. Revelation

4:11). It is impossible to create a machine that has a

consciousness, because our consciousness (and therefore

intelligence) is given by God.

I do not believe that man is capable of creating a truly

intelligent being. Proverbs 2:6 says, “For the Lord giveth wisdom: out of

his mouth cometh knowledge and understanding.” God in his power is the

only one that is capable of giving wisdom and knowledge. I

believe as humans we cannot, and should not attempt to create an

intelligent being, for only God alone is able to do this.

Intelligence comes from the Lord, and the purpose of being able

to reason should be for the glory God (The fear of the Lord is the

beginning of knowledge: but fools despise wisdom and instruction- Proverbs 1:7).

I am not opposed to technological innovation, and encourage

further development of technology that will increase the quality

24

of life that we enjoy today. However, the concept of general AI

supported by scientists (Transhumanists) like Ray Kurzweil is

flawed and contradicts Christian belief. Why would man be so

interested in the possibility of a technological singularity?

Genesis 11:4 shows man’s intention in his heart, “…let us make us a

name.” The story of babel shows us that, in his pride, man

rebelled against the Lord’s will. In our sinful nature we seek

our own desires, and advancement of personal agendas. Psalm 10:3

says that the wicked “boasteth of his heart's desire, and blesseth the covetous,

whom the Lord abhorreth.” The singularity approach to general

intelligence, seeks to create a godless world, were we are our

own god’s and control our destiny.

If the singularity were possible, we would be able to

achieve immortality by transferring our consciousness and

physical bodies to technological entities.42 As sinners our

penalty is death both physical and spiritual. Our bodies are

going to pass away someday, and the only solution to eternal

spiritual death is salvation through Jesus Christ (For the wages of

sin is death; but the gift of God is eternal life through Jesus Christ our Lord- Romans

6:23). The main belief and goal of singularity supporters is

25

based on the need to find an answer to death, which is the

penalty for our sin.

The pursuit of enlightenment and knowledge is not being done

with the purpose of glorifying the Lord, but rather to glorify

ourselves as supreme beings. This is not only the wrong approach

to take, but in the end is also a fruitless endeavor. As Solomon

says in Ecclesiastes 2:11, “Then I looked on all the works that my hands had

wrought, and on the labour that I had laboured to do: and, behold, all was vanity and

vexation of spirit, and there was no profit under the sun.” For these reasons I

concur that strong AI is not possible; even if the idea of a

strong AI agent can be entertained, I have also concluded that

this pursuit is not done with the right intentions. In the years

to come, we will be able to see new developments in the AI field,

and the attempts of different researchers to achieve this general

intelligence. For future developments of AI, I believe that

research should be directed towards development of weak AI, and

the idea of strong AI should not be considered. There are great

achievements and technological advancements that can be made

without the need for a singularity approach.

Bibliography

26

AI Magazine. “A (Very) Brief History of Artificial Intelligence.”Association for the Advancement of Artificial Intelligence. Accessed November 9, 2014. http://www.aaai.org/ojs/index.php/aimagazine/article/view/1848/1746. 53-60.

Aitopics. “Brief History.” Association for the Advancement of Artificial Intelligence. http://aitopics.org/misc/brief-history.

Alan Turing. “Computing Machinery and Intelligence.” Association for the Advancement of Artificial Intelligence. Accessed November 9, 2014.http://aitopics.org/sites/default/files/classic/Feigenbaum_Feldman/Computers_And_Thought_Turing.pdf. 11-35.

Anderson, Michael, and Susan Leigh Anderson. "Machine Ethics: Creating an Ethical Intelligent Agent." AI Magazine 28, no. 4 (Winter2007 2007): 15-26. Academic Search Elite, EBSCOhost (accessed November 24, 2014).

Bedau, Mark A., et al. "Open Problems in Artificial Life." Artificial Life 6, no. 4 (Fall2000 2000): 363-376. Academic Search Elite, EBSCOhost (accessed November 23, 2014).

Bishop, Mark. "Why Computers Can’t Feel Pain." Minds & Machines 19, no. 4 (November 2009): 507-516. Academic Search Elite, EBSCOhost (accessed November 20, 2014).

Burgin, Mark. "How We Know What Technology Can Do." CommunicationsOf The ACM 44, no. 11 (November 2001): 82-88. Academic Search Elite, EBSCOhost (accessed November 1, 2014).

Cadwalladr, Carole. “Are the robots about to rise? Google's new director of engineering thinks so.” The Guardian News. Accessed November 7, 2014. http://www.theguardian.com/technology/2014/feb/22/robots-

27

google-ray-kurzweil-terminator-singularity-artificial-intelligence.

Campa, Riccardo. "Technological Growth and Unemployment: A GlobalScenario Analysis." Journal Of Evolution & Technology 24, no. 1 (February 2014): 86-103. Academic Search Elite, EBSCOhost (accessed November 20, 2014).

Chopra, Samir. 2010. "Rights for Autonomous Artificial Agents?" Communications Of The ACM 53, no. 8: 38-40.

Cole-Turner, Ronald. "The Singularity and the Rapture: Transhumanist and Popular Christian Views of the Future.” Zygon: Journal Of Religion & Science 47, no. 4 (December 2012): 777-796. Academic Search Elite, EBSCOhost(accessed November 29, 2014).

Cordeiro, Jose. "The singularity is nigh." Engineering & Technology (17509637) 5, no. 1 (January 23, 2010): 27-29. Academic Search Elite, EBSCOhost (accessed November 15, 2014).

Cordeschi, Roberto. “AI Turns Fifty: Revisiting Its Origins.” PhilPapers. Accessed November 3, 2014. http://philpapers.org/rec/CORATF

Dean, Thomas, James Allen, and Yiannis Aloimonos. Artificial Intelligence: Theory and Practice. Redwood, CA: The Benjamin/Cummings Publishing Company, Inc., 1995.

Del Monte, Louis A. The Artificial Intelligence Revolution: Will Artificial Intelligence Serve us or Replace us? USA: Louis A. Del Monte, 2013.

Dreyfus, Hubert. "A History of First Step Fallacies." Minds & Machines 22, no. 2 (May 2012): 87-99.

Freedman, David H. Brainmakers. New York, NY: Simon & Schuster, 1994.

28

Galán, Severino F., Ole J. Mengshoel, and Rafael Pinter. "A NovelMating Approach for Genetic Algorithms." Evolutionary Computation 21, no. 2 (Summer2013 2013): 197-229. Academic Search Elite, EBSCOhost (accessed November 11, 2014).

Geraci, Robert M. "Robots and the Sacred in Science and Science Fiction: Theological Implications of Artificial Intelligence." Zygon: Journal Of Religion & Science 42, no. 4 (December 2007): 961-980.Academic Search Elite, EBSCOhost (accessed November 15, 2014).

Golubev, Yu. "Neural networks in mechatronics." Journal Of Mathematical Sciences 147, no. 2 (November 22, 2007): 6607-6622. Academic Search Elite, EBSCOhost (accessed November 19, 2014).

Guarino, Alessandro. "Intelligent weapons are coming Sj." Engineering & Technology (17509637) 8, no. 8 (September 2013): 75-77. Academic Search Elite, EBSCOhost (accessed November 22, 2014).

Hall, J. Storrs. Beyond AI. Amherst, NY: Prometheus Books, 2007.

Hayes, Patrick I., and Leora Morgens Tern. "On John McCarthy's 80th Birthday, in Honor of His Contributions." AI Magazine 28, no. 4 (Winter2007 2007): 93-102. Academic Search Complete, EBSCOhost (accessed November 16, 2014).

Hearst, Marti, and Haym Hirsh. “AI’s greatest trends and controversies.” Cornell University, Department of Computer Science. Accessed November 9, 2014. http://www.cs.cornell.edu/courses/cs472/2002fa/handouts/challenges-ai.pdf.

Herzfeld, Noreen. "Creating in Our Own Image: Artificial Intelligence and the Image of God." Zygon: Journal Of Religion & Science 37, no. 2 (June 2002): 303. Academic Search Elite, EBSCOhost(accessed November 29, 2014).

29

Hillis, Danny, et al. "In Honor of Marvin Minsky's Contributions on his 80th Birthday." AI Magazine 28, no. 4 (Winter2007 2007): 103-110. Academic Search Elite, EBSCOhost (accessedNovember 23, 2014).

Hogan, James P. Mind Matters: Exploring the World of Artificial Intelligence. New York, NY: The Ballantine Publishing Group, 1998.

Hongo, Jun. “Artificial Intelligence Outperforms Average High School Senior.” Dow Jones & Company, Inc. http://blogs.wsj.com/japanrealtime/2014/11/04/artificial-intelligence-outperforms-average-japanese-high-school-senior/.

IEP. “Artificial Intelligence.” Internet Encyclopedia of Philosophy. http://www.iep.utm.edu/art-inte/#SH3b.

Isaac Asimov. I, Robot. New York, NY: Doubleday & Company, 1950.

Jackelén, Antje. "The Image of God as Techno Sapiens." Zygon: Journal Of Religion & Science 37, no. 2 (June 2002): 289. AcademicSearch Elite, EBSCOhost (accessed November 15, 2014).

Kenaw, Setargew. "Hubert L. Dreyfus’s Critique of Classical AI and its Rationalist Assumptions." Minds & Machines 18, no. 2 (May 2008): 227-238. Academic Search Elite, EBSCOhost (accessed November 15, 2014).

Kurzweil, Ray. "The Future of Intelligent Technology and Its Impact on Disabilities." Journal Of Visual Impairment & Blindness 97, no. 10 (October 2003): 582-584. Academic Search Elite, EBSCOhost(accessed November 15, 2014).

Legg, Shane, and Marcus Hutter. "Universal Intelligence: A Definition of Machine Intelligence." Minds & Machines 17, no.4 (November 2007): 391-444. Academic Search Elite, EBSCOhost (accessed November 10, 2014).

30

McCorduck, Pamela. Machines Who Think. Natick, MA: AK Peters, Ltd, 2004.

McGinnis, John O. "Accelerating AI." Northwestern University Law Review 104, no. 3 (Summer2010 2010): 1253-1269. Academic Search Elite, EBSCOhost (accessed November 5, 2014).

Muehlhauser, Luke, and Bill Hibbard. "Exploratory Engineering in Artificial Intelligence." Communications Of The ACM 57, no. 9 (September 2014): 32-34.

Nasso, Christine. Artificial Intelligence: Opposing Viewpoints. Farmington Hills, MI: Greenhaven Press, 2011.

Neate, Roger. "Free Speech for Algorithms?." Communications Of The ACM 54, no. 3 (March 2011): 6. Business Source Elite, EBSCOhost (accessed November 11, 2014).

Prudkov, Pavel N. "A View on Human Goal-Directed Activity and theConstruction of Artificial Intelligence." Minds & Machines 20, no. 3 (August 2010): 363-383. Academic Search Elite, EBSCOhost (accessed November 1, 2014).

Shyamasundar, R. K. "The computing legacy of Alan M. Turing (1912-1954)." Current Science (00113891) 106, no. 12 (June 25,2014): 1669-1680. Academic Search Elite, EBSCOhost (accessed November 23, 2014).

Stanford Encyclopedia of Philosophy. “The Chinese Room Argument.”Stanford University. http://plato.stanford.edu/entries/chinese-room/.

State of California. “Senate Bill No. 1298.” California Legislation. http://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201120120SB1298.

Szoldra, Paul. “Elon Musk Thinks Sci-Fi Nightmare Scenarios AboutArtificial Intelligence Could Really Happen.” Business

31

Insider. Accessed November 5, 2014. http://www.businessinsider.com/elon-musk-artificial-intelligence-mit-2014-10.

Tonkens, Ryan. "A Challenge for Machine Ethics." Minds & Machines 19, no. 3 (August 2009): 421-438. Academic Search Elite, EBSCOhost (accessed November 1, 2014).

US Congress. “Cybersecurity Information Sharing Act of 2014.” Library of Congress. https://www.congress.gov/bill/113th-congress/senate-bill/2588.

Vardi, Moshe Y. "Who Begat Computing?." Communications Of The ACM 56, no. 1 (January 2013): 5. Business Source Elite, EBSCOhost (accessed November 17, 2014).

Vladeck, David C. "Machines Without Principals: Liability Rules and Artificial Intelligence." Washington Law Review 89, no. 1 (March 2014): 117-150. Academic Search Elite, EBSCOhost (accessed November 23, 2014).

Warwick, Kevin, Huma Shah, and James Moor. "Some Implications of a Sample of Practical Turing Tests." Minds & Machines 23, no.2 (May 2013): 163-177.

Weber, Karsten. "What is it like to encounter an autonomous artificial agent?." AI & Society 28, no. 4 (December 2013): 483-489. Academic Search Elite, EBSCOhost (accessed November 1, 2014).

Whitby, Blay. Artificial Intelligence: A Beginner’s Guide. Oxford, England: OneWorld Publications, 2003.

Appendix

Appendix A

Biographical Information

32

Alan Turing: One of the leading scientists for deciphering the German Enigma code in the early 1940s. The first scientist towrite a philosophical analysis of the possibility of AI in his 1950 paper Computing Machinery and Intelligence.43 Devised the Turing Test in 1950. Which is still used today in annual competitions to measure the ability of machines to communicate like a real human being.44

John McCarthy: Born in Boston in 1927. Attended the California Institute of Technology and obtained a bachelor and graduate of science.45 McCarthy obtained a PhD from Princeton in1950. In 1952 he joined the faculty of mathematics in Dartmouth College. It is here in Dartmouth that McCarthy held the first conference and started formal research in the AI field.46 In 1963McCarthy invented the LISP language for AI programming. This yearMcCarthy also began his lifelong research project, which is to create a computer that possesses common sense.47

Appendix B

Asimov’s three laws of robotics were proposed in his fictionbook, I Robot. These laws were proposed by Asimov as a rule for creating AI that is friendly towards the human race:

1. A robot may not injure a human being, or through inactionallow a human being to come to harm.

2. A robot must obey the order given it by human beings except where such order would conflict with the first law.

3. A robot must protect its own existence as long as such protection does not conflict with the first or second law.48

33

Appendix C

The following conversation held between the AI agent ELIZA, and a person shows the ability that symbolic symbols have to analyze key words and provide appropriate responses. Although ELIZA is an older system, it still proved that computers in a waymimic human intelligent behavior.

ELIZA: I am the psychotherapist. Please, describe your problems.Person: Hello, can you help me?ELIZA: How do you do? What brings you to see me?Person: My wife asked me to see you.ELIZA: Tell me something about your family.Person: She says I use the computer too much.ELIZA: You have your mind on computers, it seems.Person: Yes, and my hands, too.ELIZA: Why do you say yes and your hands too?Person: Because I use it all the time.ELIZA: Is it because you use it all the time that you cameto me?49

Appendix D

Basic structure of an expert system (Knowledge-Based systems):50

Working Memory

Inference Engine

Interface

Knowledge Base

34

The Knowledge Base is data input by specialized experts to the system. Working Memory is the information the system (such as MYCIN) would use for the specific case the user is inquiring about. The inference engine would be the algorithm that sorts thedata input from the working memory and tries to match the information with the stored knowledge base. The interface allows the user to interact with the expert system and allows a user to reevaluate the question and receive appropriate responses.51

Appendix E

Bill 1298, passed in California (2012), is one of the examples of legislation made by government on a technological product produced by the efforts of the AI field.52

Senate Bill No. 1298

CHAPTER 570

An act to add Division 16.6 (commencing with Section 38750) tothe Vehicle Code, relating to vehicles.

[Approved by Governor September 25, 2012. Filed with Secretary ofState September 25, 2012. ]

LEGISLATIVE COUNSEL'S DIGEST

User

35

SB 1298, Padilla. Vehicles: autonomous vehicles: safety andperformance requirements.

Existing law requires the Department of the California HighwayPatrol to adopt rules and regulations that are designed topromote the safe operation of specific vehicles, including, amongother things, schoolbuses and commercial motor vehicles. Existinglaw also requires the Department of Motor Vehicles to registervehicles that are being operated in this state and to issue alicense plate to an applicant for the operation andidentification of that person’s vehicle.

This bill would authorize the operation of an autonomous vehicle,as defined, on public roads for testing purposes, by a driver whopossesses the proper class of license for the type of vehiclebeing operated if specified requirements are met, including thatthe driver be seated in the driver’s seat, monitoring the safeoperation of the autonomous vehicle, and capable of taking overimmediate manual control of the autonomous vehicle in the eventof an autonomous technology failure or other emergency. The billwould prohibit, except as provided for testing purposes, theoperation of such a vehicle on public roads until themanufacturer submits an application to the department thatincludes various certifications, including a certification thatthe autonomous technology satisfies certain requirements, and theapplication is approved by the department pursuant to theregulations that the department would be required to adopt. Thebill would require one of the certifications to specify that theautonomous vehicle’s technology meets Federal Motor VehicleSafety Standards for the vehicle’s model year and all otherapplicable safety standards and performance requirements setforth in state and federal law and the regulations promulgatedpursuant to those laws.

The bill would require that the Department of Motor Vehiclesadopt regulations as soon as practicable, but no later thanJanuary 1, 2015, setting forth requirements for the submission ofevidence of insurance, surety bond, or self-insurance required bythe bill and requirements for the submission or approval of anapplication to operate an autonomous vehicle, including any

36

testing, equipment, or performance standards, as specified, andto hold public hearings on the adoption of any regulationapplicable to the operation of an autonomous vehicle without thepresence of a driver inside the vehicle. The bill would providethat federal regulations promulgated by the National HighwayTraffic Safety Administration supersede state law or regulationwhen found to be in conflict.

The bill would require the department to approve an applicationsubmitted by a manufacturer upon making specified findings andwould authorize the department to impose additional requirementsif the application seeks approval for autonomous vehicles wherethere is no person in the driver’s seat. The bill would alsorequire the department to notify the Legislature of the receiptof an application from a manufacturer seeking approval to operatean autonomous vehicle capable of operating without the presenceof a driver inside the vehicle and the approval of theapplication. The bill would provide that approval of theapplication is effective no sooner than 180 days after the datethe application is submitted.

The department would be authorized to charge a fee for theapplication in an amount necessary to recover all costsreasonably incurred by the department.

37

End Notes

1 Jun Hongo. “Artificial Intelligence Outperforms Average High School Senior.” Dow Jones & Company, Inc.

2 Louis A. Del Monte. The Artificial Intelligence Revolution: Will Artificial Intelligence Serve us or Replace us? (USA: Louis A. Del Monte, 2013), 33.

3 James P. Hogan. Mind Matters: Exploring the World of Artificial Intelligence. (New York, NY: The Ballantine Publishing Group, 1998) 70.

4 Mark Bishop. "Why Computers Can’t Feel Pain." Minds & Machines 19, no. 4 (November 2009): 507.

5 Louis A. Del Monte. The Artificial Intelligence Revolution: Will Artificial Intelligence Serve us or Replace us? (USA: Louis A. Del Monte, 2013), 25.

6 Pamela McCorduck. Machines Who Think. (Natick, MA: AK Peters, Ltd. 2004), 19.

7 Isaac Asimov. I, Robot (New York, NY: Doubleday & Company, 1950), 85.

8 Aitopics. “Brief History.” Association for the Advancement of Artificial Intelligence.

9 Alan Turing. “Computing Machinery and Intelligence.” Association for the Advancement of Artificial Intelligence.

10 Aitopics. “Brief History.” Association for the Advancement of Artificial Intelligence.

11 Ibid.12 Pamela McCorduck. Machines Who Think. (Natick, MA: AK Peters,

Ltd. 2004), 442.

13 Aitopics. “Brief History.” Association for the Advancement of Artificial Intelligence.

14 Louis A. Del Monte. The Artificial Intelligence Revolution: Will Artificial Intelligence Serve us or Replace us? (USA: Louis A. Del Monte, 2013), 43.

15 Kenaw, Setargew. "Hubert L. Dreyfus’s Critique of Classical AIand its Rationalist Assumptions." Minds & Machines 18, no. 2 (May2008): 228.

16 Aitopics. “Brief History.” Association for the Advancement of Artificial Intelligence.

17 Hall, J. Storrs. Beyond AI. (Amherst, NY: Prometheus Books, 2007), 65.

18 Aitopics. “Brief History.” Association for the Advancement of Artificial Intelligence.

19 James P. Hogan. Mind Matters: Exploring the World of Artificial Intelligence. (New York, NY: The Ballantine Publishing Group, 1998), 246.

20 Louis A. Del Monte. The Artificial Intelligence Revolution: Will Artificial Intelligence Serve us or Replace us? (USA: Louis A. Del Monte, 2013), 17.

21 Aitopics. “Brief History.” Association for the Advancement of Artificial Intelligence.

22 Yu Golubev. "Neural networks in mechatronics." Journal Of Mathematical Sciences 147, no. 2 (November 22, 2007): 6607.

23 Ibid, 6007.

24 Blay Whitby. Artificial Intelligence: A Beginner’s Guide. Oxford, England: One World Publications, 2003.

25 James P. Hogan. Mind Matters: Exploring the World of Artificial Intelligence. (New York, NY: The Ballantine Publishing Group, 1998), 183.

26 Carole Cadwalladr. “Are the robots about to rise? Google's newdirector of engineering thinks so.” The Guardian News.

27 Blay Whitby. Artificial Intelligence: A Beginner’s Guide. Oxford, England: One World Publications, 2003.

28 Ibid.29 Hubert Dreyfus. "A History of First Step Fallacies." Minds &

Machines 22, no. 2 (May 2012): 90-91.

30 Szoldra, Paul. “Elon Musk Thinks Sci-Fi Nightmare Scenarios About Artificial Intelligence Could Really Happen.” Business Insider.

31 Anderson, Michael, and Susan Leigh Anderson. "Machine Ethics: Creating an Ethical Intelligent Agent." AI Magazine 28, no. 4 (Winter2007 2007): 15.

32 John O. McGinnis. "Accelerating AI." Northwestern University Law Review 104, no. 3 (Summer2010 2010): 1260.

33 Ibid, 1259.

34 Ibid, 1254.

35 AI Magazine. “A (Very) Brief History of Artificial Intelligence.” Association for the Advancement of Artificial Intelligence.

36 Ibid.

37 Riccardo Campa. "Technological Growth and Unemployment: A Global Scenario Analysis." Journal Of Evolution & Technology 24, no. 1 (February 2014): 87-88.

38 Ibid, 88.

39 Ibid, 87.

40 Ibid, 88.

41 Alan Turing. “Computing Machinery and Intelligence.” Association for the Advancement of Artificial Intelligence. 13.

42 Ronald Cole-Turner. "The Singularity and the Rapture: Transhumanistand Popular Christian Views of the Future.” Zygon: Journal Of Religion & Science 47, no. 4 (December 2012): 790.

43 R. K. Shyamasundar. "The computing legacy of Alan M. Turing (1912-1954)." Current Science (00113891) 106, no. 12 (June 25, 2014): 1669.

44 Ibid, 1679.

45 Hayes, Patrick I., and Leora Morgens Tern. "On John McCarthy's80th Birthday, in Honor of His Contributions." AI Magazine 28, no. 4 (Winter2007 2007): 94.

46 Ibid, 95.

47 Ibid, 96.

48 Pamela McCorduck. Machines Who Think (Natick, MA: AK Peters, Ltd, 2004), 25.

49 Hall J. Storrs. Beyond AI (Amherst, NY: Prometheus Books, 2007), 65.

50 James P. Hogan. Mind Matters: Exploring the World of Artificial Intelligence (New York, NY: The Ballantine Publishing Group, 1998), 246.

51 Ibid, 247.

52 State of California. “Senate Bill No. 1298.” California Legislation. http://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201120120SB1298.