L1CHAPTER

INTRODUCTION TO Al .

Ill INTRODUCTION
We ne~d to de~elop some paradigms or algorithms that cause our machines to ?e~orm _tas~ that
otherwise requrre cognition or perception when perfonned by humans. Any art1fic1ally mtelhgent
syStem must possess ti?,ree-essential components:- ·
1. A means for representation of diverse kinds of knowledge
Knowledg~ may be generic and domain specific, implicit and at different level~ of abstraction. The
representation mechanism that we opt must be able to handle this knowledge ip. any form. Please
note that a representational structure enforces certai~Iimitations on the nature of inferences
that can be drawn from the embedded knowledge. ·
2. A framework for reasoning
We need to have some control mechanisms to constrain the search through a Knowledge Base (KB)
and the means of arriving at conclusions.
3. A mechanism for learning
An artificial intelligent system must have a method of learning new data, storing it in the existing
structures internally with minimal or no disturbances to them.
Classical AI or Good Old Fashioned AI is symbolic and top-down in its approac_h.
What is Al exactly?
According to Patterson, "Al is a branch of computer science that deals with the study and the
creation of computer systems that exhib~t some form of intelligence." By 'intelligence' we mean
(a) Systems that learn new concepts arid tasks; ·
(b) Systems that can reason and draw useful conclusions about the world around us;
(c) Systems that can understand a natural language or perceive and comprehend a visual scene; and
(d) Systems that perform. other types of feats that require human types of intelligence.
Many other definitions of Al have been given:
l . General definition:
. "An understanding
. ofAI requires\..an understanding ofrelated tennssuch
as intelligence, knowledge, reasomn~, thought, cogmtton, le~g and solving problems."
2_ By Advert: "Al i~ the part of comp_u~er scienc~ ~oncerned with designing intelligent
Computer systems i.e. systems that e~btt charactenst1cs that we associate with me
• t 11 tgence
. .
m
human behavior."
3. By heuristic (i.e., Rule of Thumb): "AI is the branch of computer science that deal .th
. . b l s w1 the
ways of representing knowledge usmg sym o s rather than numbers and with th
· infi · " e ru1es of
. thumb for processmg ormatton.
1
r
I
I
2 ARTIFICIAL INTELLIGEtJCE A PRACTICAL APPROACH

4 · Modern definition: "Al is defmed as the branch of computer science dealing wi~
11

symbolic and non-algorithmic method of problem solving."

I ~OTE: AI works with "pattern-matching' methods which attempt to describe object, events and process
\ m tenns of their qualitative nature and logical and combinationaf relationship.
I. . Please understand that "intelligence" is a catchy word here. Intelligence means an integrated
I
sum of all knowledge and facts, acquired through study and experience. Also note that the food for
this intelligen~e is knowledge.
Al task domain ·
Basically, AI covers three types of tasks:
(a) Mundane tasks
(b) . Formal tasks
(c) Expert tasks
Let us understand these tasks now.
I. Mundane tasks
AI has some mundane tasks also like,
(a) Perce~tion of vision and speech.
(b) Natu.ral laI,lguage understanding . .
(c) Common sense reasoning.
(d) Robotics.
Il. Formal tasks
AI formal tasks include:
(a) Game playing, e.g., chess, 8-queens problems, water-jug problem etc.

(b) Mathematics-geometry, calculus, etc.

m. Expert tasks
It includes:
. (a) Engineering field.
(b) Scientific analysis.
(c) Medical and financial.analysis.
What Al is N OT?
1. Al is not tlie study and creation of conventional computer systems.
All programs.exhibit some degree of intelligence and an Al program must go beyond this degree
of intelligence also. This is not the case. Remember that human beings are more intelligent than our
intelligent systems also as we only have created them.
2. Al is not the study of mind, nor of body, nor oflanguages, as found in_fields of psychology
physiology, cognitive science or linguistics.
Although there is some overlap between these.fields and AI, yet the goal of Al. is to develop a
computer system that i,s capable of performing intelligent tasks effectively and efficiently.
So, what AI will include? It includes areas like:
(a) Robotics
(b) Memory organization
(c) Knowledge representation
 INTRODUCTION TO Al 3
(d) Storage and recall
(e) Learning models
(j) Inferenc~ techniques
(g) Common sense reasoning
(h) Decision making.
(i) Pattern recognition
(j) Searching
(k) Speech recognition
(l) Speech synthesis.
History of Al
Let us see the history
., of AI in 1950s,
. 1960s' 1970s, 1980s and 1990s
· m . a tabular form now.
Year , · ·. · .Developments mad~
1950s O Birth of AI
O. First Neural Net Simulator (Minsky)
O GPS-General ~ose Problem Solver (Simon)
O GTP-Geometry . Theorem prover-(Gelertber), mput
. diagrams back-
ward reasonmg. '
o SAINT-Symbolic integratio~ ..
1960s O ~NALOGY-could solve IQ test puzzle
O STUDENT-could solve algebraic word problems
O SHRDLU-could manipulate blocks using robotic arm
o STRIPS-_a problem solver planner
O Minsky and Papert-de_ monstrated the limitations of neural nets

1970s o Conceptual Dependency theory (Shank)

o Frames{Minsky) ··
o '
Machine Learning: ID3 (Quinlan), AM (Lenat)
o Expert system got success
O DENDRAL by Feigenbaum- to identify molecular structure.
O MYCIN by Shortl1ffe and Buchanan- to diagnose infectious blood diseases
O Cheaper computing, so AI Software a feasible success, neural nets revisited
· 1980s
· o -XCON (by McSermott}--saved $ 40m per year ·
· o Neural computing-back propagati9n (Werbos), associative memory
(Hopfield), logic _programming developed an~ used
1990s O Embedded intelligent systems, agents
o Hybrid approaches: logic+ neural nets+ genetic algorithms+ fuzzy logic
o CYC (by Lenat}--a far-reaching project to capture common sense
reasoning
o Society of Mind (Minsky}--Intelligence is a product of complex
interactions of simple agents
o Deep Blue (formerly Deep Thought}--defeated Kasparov in speed chess
in 1997
 4 ARTIFICIAL INTELLIGENCE -A PRACTICAL APPROACH

Let us see the entir'e h'lStory now, from 1941 to 1991. '
Al-based .hardware
The birth of Al sells $425 million
Start of DOD's to companies
advanced research
projects Al system beat~
I human chess
First'
Electronic Dartmouth First expert system
m•r·
computer
conference
I
I I 1968 1970 1972
1986 1991

l ..
1958 1963
1949 1956
1941 Al military Jystems
-, .
Micro world
used effectively in
Lisp language program SHRDLU DESERT STORM
I
developed created ·
First commercial
stoted program
computer
PROLOG language
Logic theorist revealed
developed

Fig. 1.1 History of Al.

Ill SCOPE OF ~I .
AI bas a very wide scope. Some of its applications are discussed below:

1.1.1. Games
According to NeweJI and Simon (1976), the <Ssential basis f~r human probl~m solvin~ is to
• systematically explore a space of problem states, i.e., succe.,1ve and alternattve stages m the
problem solving, For example, the different board configura1tons ma chess game or mtennedtate
steps in a reasoning process. This space is called as state space. This space of alternattve soluttons
is then searched to find a final answer.
Please understand that games can generate extre)'1ely large search spaces, So, we nccd
powerful tecJmiques to search for our solution in t!tl• space. These techniques are called as
heuristics. Also note that intelligence resides in,lteuristics only, So, Al helps here.

1.1.2 Theorem Proving

It is a fonnal system wherein logic lends itself to automation. A large variety of problems can be
solved by using logical axioms and treating problem instances as theorems to be proved. Theorem
proving has devised powerful solution heuristics and reduced the complexity of the search space.
For example, design and verification oflogic circuits, control of complex systems etc., will respond
to such an approach.
Many modem theorem provers function as 'intelligent assistants'. By this we mean that it
allows.humans to do the task of decomposition of large problems into subproblems and devising
h;unsttcs for searching the space of possible proofs. On the other hand, the theorem prover then
P rfonns the srmpler but sttll demandmg task of proving lemmas, verifying smaller guesses and
completmg the fonnaJ aspects of a proof outlined by the humans.
INTRODUCTION TO Al
1.1.3
.
N atural Lan guage Processing (NLP) 5
It 1s a subfield of Al h.
ownd natural
b la nguage.
w •ch
Thisdeals
will with
fill ththe methods of commumcaliDg
. . with . a <'omputer in one's
nee not e computer literate to comm e_gap between the humans and the machines. So, now one
1 4 v· . . urucate with it
1·. . 1s1on and Spee h P .
.. c rocessing
ComputerIvision. is a comput atlon
. mten
. . .
more on c assical AI methods f symbohc si:eprocess·
process. It involves multiple, transfionnat10ns.
• Thi s rehes
.
O

. Speech. understanding
· . recogru·f
requrres f mg.
against
· ·1ex1con patterns for recogrut1on
. . Devion Io basic
• speech patterns· These patterns are matched
contmumg goa of AI researchers. . e opmg systems that understand speech has been a
1
1.1.5 Robotics
Accordingti to the Robot
• Institute 0f . (RIA)
mu unc onaI, mampular that is d Amenca
. - "A robot is a - reprogrammabl
ltif
. th h .
eVIces roug vanous programmed es1gned
. to ·. move matenals,
. parts, tools or specializede,
d motions for the perf
Today, we need intelligent Robots to ak h ormance of a variety of tasks."
· · · m e t em
embe e mto 1t. So, now vision is very e t· Th . evices. smaller d · so, mtelligence
· needs to be
·dd d • ssen ia1 ese i t ll'
to solve our complex problems easily. . · n e ,gent robots as well as Al promise u.s

1.1.6 Expert Systems (ES)

. knowledge
They are h intensive
. . programs that solve problems m
· a domam
· · that needs a d t hni 1
expertise. As per t e defimt10n given by 'British Computer Society's Co . goo ec . ~a
Group on Expert Systems- - - m~ee_ofthe_Specialist
"The
· embodiment
h ti within a compµcer
. of a. knowledg.e-based
. componen
· t. _!om-an
f -" -·expert
s ID sue a orm that
· ti the machme.,can
· Ad . offer mtell1gent advice or take an intelligent dec1S1on
·. ·
kill
a out a processing unction. e~ble additional characteristic which many would ;. ard
b
fundame~tal, is ~• c•!'~bility of t~e sy~tem on demand to justify its own line of reasonin:in a
manner directly mtelbgible to the mqu1rer. The style adopted to attain these characteristics is
rule-based programming."
So, we find from this dOfinition that an Al specialist or a knowledge engineer or expert system
·designers are responsible for implementing knowledge in a program, both effectively and intelligently.

For example,
1. DEN DRAL - developed at Stanford University in late 1960s. It was de.signed to infer
the structure of mganic mo_leculOs from their chemical formulas and-mass spectrographic
information about the chemical bonds present in the molecules. As thes~oigami;molecules
tend to be very large, so the number of possible structures for these molecules tend to be
huge. DENDRAL addresses this problem of a large search space by applymg the beunsllcs
knowledge of expert chemists to the structure elucidation proolem.
DENDRAL used domain specific knowledge for this.
2. M\'CIN -It uses the expert medical knowledge to diagnose and prescribe trea11nentrd for ·
spinal meningitis and bacterialillfections of the blood. It was developed at Stanfo m mid·
1970s. It provides cleai and logical explanations of its reasoning. It uses a control SlrllCture
appropriate to the specific problem domain.
3. · PROSPECTOR - A program for determining the probable location and the type of ore
deposits based on geological information
' about a site.
. I f1vely specialized expert
Please note that most expert systems have been wn tten ,or re a
level domains.
 6
ARTIFICIAL INTELLIGENCE A PRACTICAL

- Al TECHNIQUES
I. Search Knowledge
Knowledge can be defined as the body of facts and principles accumulated by bu .·
the act, fact or state of knowing. For example, in biological organisms, knowledge~ankind
comp Iex structures of mterconnected
· neurons. The structures correspond to symbolic repr1s stored ~
of the knowledge possessed by the organism, the facts, rules and so on. Please note that ;:entati'
human brain weighs about 3.3 pounds and ·contains an estimated number of 1012 navera,
Also note that these neurons and their interconnection capabilities provide about 1:~r~~
of.potential storage capacity. On the other hand, in computers knowledge is .stored as SYJnb
01
structures but in form of collections of magnetic spots and voltage states. ~
Knowledge is of three types as shown below:

I Knowledge
I
I I I
Procedural Declarative .
or or Heuristic
Operational Relational knowledge
knowledge · knowledge

. Fig. 1.2 Types of knowledge

1

. Let us defin~ these types now.

1. Procedural or Operational Knowledge
It is defined as a compiled knowledge ~elated to the performance of some task. For.example, steps
to solve a quadratic equation are expressed as·a procedural knowledge.
2. Declarative or Relational Knowledge ·
It is apassive knowledge expressed as statements of facts ;ibout the world. For example, personnel
data in a database. Such data are explicit pieces of independent knowledge. ·
3. Heuristic Knowledge
· Heuristics means using some rules of thumb or tricks or some str~tegies to simplify the solution to
problems. We acquire this after much experience.
Some terminologies related to knowledge
We define certain terms that will be used again and again here. They are as follows:
1. Knowledge and data: A doctor has both knowledge as well as data. Here, data is the ·
patient's record whereas knowledge is what he has learned in his medical college. This was
explained by Feigenbaum and McCorduck.
2. Belief: It is defined as essentially any meaningful and coherent expression that can be
represented. It may be true or false.
3. Hypothesis: It is defined as a justified belief that is not known to be true. It is backed up
with sonie supporting evidences but it may still be false. ·
4. Knowledge: It is a true justified belie£
5. Meta knowledge: It is the knowledge about the knowledge.
6. Epistemology: It is the study of the nature of knowledge.
Al technique is a method that exploits knowledge that should be represented in such a way that:
I. It captures generalizations.
2. It can be understood by people who must provide it.
 r10N roAL
,Nr~ootJCcan be·east·iy modified to. correct errors
. . .
and to inc
orporate cha 7
3. It d manY situations even 1f 1t 1s not totally nges.
It can be use . accurate or
4. b used to help overcome its own sheer hulk b . complete.
5. It c~bl:ies that roust usually be considered. Yhelpmg to nan
poss1 11 ow the ran
hat it is possible to solve AI problems with . ge of
I se note t Als ·t . . out us1n AI
Pea Id be inefficient. o 1 1s possible to apply AI g techn·
soI•ti••5 "'~edlS and this will · be a good thing ·for those techniqnes to ~~:: h~t tbe
non-Ai P~ti , as Al problems have. problems that poss •Hon ol
baracter1s c . . . . ess same
c stelilS that depend on a rich base of knowledge t
those sy o perform d'
owledge-based systems. ifficult tasks
1a1own as kn . are
A owledge-based system (K.BS) consists of three main comp onents as shown m . Fig.
. 1.3.
kn
__
-
-
.
Query lnupt-Output .. Inference
Control Unit . Knowledge
Unit Base
-
Fig. 1.3 Components of a KBS
Explanation:
·· dK.BS-s dget their
Thpower
kn from
l d the
· expert knowledge that has been coded mto. facts
rules, heuri~tics anl/ pro~e urdelns.fi e oCwUe) ge ts stored~ a knowledge base only. Since it is stored
separately (from . O. umtknan l derence 'th t so we can
·1· eastly add new knowledge to this knowledge
.
base or refine ex1stmg ow e ge WI . ou recomp1 mg the control
• and inferencm·g programs. So,
now the construct10n as well as the mamtenance of KBS bec?mes very simple.
Knowledge-base is different from datab~e. Let us tabulate the differences between them now.

,/n:ata Base,(DB) , -, ,'.

. ·Knowledge Base (KB)
,. t>\ ~'
1. It has information at a higher level of
1. It is defined as a collection of data
abstraction.
representing facts.
2. It is smaller than a DB.
2. It is larger than a KB.
3. Changes are gradual.
3. Changes are _fast. 4. It has the power of inferencing.
4. All information needs to be stated explicitly.
5. It is used for data analysis and planning.
5. u·is' mainttained for. operational purposes. 6. Knowledge is represented by logic or
6. Knowledge is represented by relational, ·
rules.
network or hierarchical model.
Knowledge may be represented at various levels as shown in Fig. 1.4.
Mental image

Written text

Character string

Binary numbers

Magnetic spots
f10n
Fig. 1.4 Levels of knowledge representa
r 8 ARTIFICIAL INTELLIGENCE -A PRACTICAL APPROAC!i
Please note, however, that any choice of representation will depend on the type of proble1t1
to be solved and the inference methods avaUable. .
Not only this, knowledge organization in memory is also very vital. The KBS discussed hi
Fig. 1.3 may require several (tens of thousands) of facts and rules to perform their intended tasks.
So, now it is mandatory to easily search, locate and retrieve appropriate facts and rules from this fOI
KB. We use indexings to speed up searches - we can form groups of knowledge of similar type (a)
and then make a pointer to point to this group. So, how instead of searching for an entire KB, only 1t i
a small portion of if may be accessed. · .
(bl
We need to manipulate this lmowledge also. Usually, the user gives some input and thus,
initiates a search for a goal or decision. Please understand that this requires that known facts lt
in the KB be l~cated, compared (matched) and altered somehow. Also note that this process (c·
may set up some sub-goals which requires some more inputs and so on until a final solution It
is found. This requires a form of inference or deduction using the knowledge and inferring rules. (~
Please note here that all forms of reasoning require certain amount of searching and matching. Also
lt
note that searching and matching both consume the maximum time of computation in AI systems.
So, we need to have the best. searching technique today so that we can avoid this combinatorial (4
explosion problem during searching. 11
Now, the question is how to acquire knowledge? Knowledge ~ay be acquired from sources like (
~:xtbooks, _refere~ces,reports, technical research papers and so on and to be useful, it should be accurate,
. mplete, mcons1stent and so on. Our KBS depends on a high quality lmowledge for their success.

Ill ALAN TUR.ING MACHINE

In 1950 Alan Turin · bli h d : . · -
topic ' "C g ?u · s e an article m the "Mind Magazine" which triggered a controversial
- an machines think?" So the Tu · t 't • "A ·
enough to have . · . ' nng es is - computer is programmed well
a conversation with an interrogato d . h
cannot discern if the . r an passes t e test if the interrogator
I . re is a computer or human at the other end."
t means that m this imitation game there are three actors -
(a) a male, ·
(b) a ferp.al~, and
(c) an interrogator. ·
That is,
Human "

or . Room'---A Human
m~chine Room-B or
machine
I I
I

Room-c

Interrogator

Fig. 1.5 Turing test

Tu ring proposed that if a human interrogator in Room C . .
who is in Room-A or Rooni-B then th h" - (Fig. 1.5) 1s not able to identify
, e mac me possesses intelli T • . .
as a sufficient test for attrihutin thinki -- . . gence. unng considered this
g ng capacity to a machine. This test is not that easy as given
here because we know that humans are far mo · ·
. . . . re supenor m common sense and reasoning. But, even
today this Turing test 1s the ulttmate test that a machine must pass in order to be called as intelligent.
 9
INTRODUCTION TOAI
So, the full Tunng
-· t ·• ,
"ClUl machmes est isB - "C~onsi
. . think?" to "d~r a machine to be ~telligent we need to test that -
?, for this llllitation machines behave intelligently?" .
8 the cbmputer nee::':" above, the Turing test .;eds an operational definition of intelbgence.
For !his
O
(a) N atural Langua p
possess the following- ·
· . ge rocessing·(NLP
It 1s reqwred to commurucate
. with R ) ·
(b) Knowledge R~ oom-A or Room-B.
. . _. presentation
It 1sAutomated
(c) reqwred to store
R a· d. . e intionnabon
n retriev . . . . ·
provided before oi dunng mterrogauon.
easonmg 1
It is M
usedh.to access stored
. intionnabon
. to answer queries .and to draw new conclus .ons.
(d) ac me Learnmg _
It is needed
.. to. adapt to ne w conditions
. . and to detect and extrapol~attems.
(e) VISIOn ' '

It is, required for total T-

. unng
· · test - to recognize
· interrogator's actions above.
-
(/) Motor Control -
It'is to act upon objects as requested . . •
(g) Senses
For total Turing test, these are needed like-smell, touch, etc.
Please understand that for a program to pass this Turing test, it needs to pass and exhibit
these capabilities. ·
Limitations of Turing Test
Turing test has certain drawbacks which are given below-
! . Say, a machine bas passed the Turing test. Now, the questiortis as I!! what level of proficiency
it has achieved? We can say that it is the level of programmer only, that is, its proficiency.
In fact, a machine demonstrates the-nitelli.gence of human brains only.
2. What happens if a machine passes this Turing test, is intelligent, manipulates all formal
symbols but Jacks understan_ding.
3. Turing teSt is not reproducible, constructive and amenable to mathematical analysis.
4. It does not consider about the physical interaction with the interrogator and the env_iromnent.

Ill CHARACTERISTICS OF Al PROBLEMS

1. Almost all AI problems have a combinatorial explosion of solutions. -

For example, in case of a chess game, the number of possible positions bas been estimated
to about 3 100, wh·h·t
1c 1s oo large. So, our conventiona
. 1computers cannot handle this huge
5
solution space. Thus, we need AI programs.
· AI programs manipulate
d 1 · h symbolic
- · information,
. on the other hand conventi·onal computers
2 (or prograins) ea wit numenc processing. .
· An AI program needs a--_ 1
_:;- KB, whereas a conventional progr~ neesaargeDB.
large d
3_ An
. Al program has an ab1hty to learn.
· Conventional systems have n0 t ac h'1eved that level
4
till now.
s. An Al program
precise needs an imprecise knowledge whereas a conventional program needs a
1a10wledge.
 10 ARTIFICIAL INTELLIGENCE A PRACTICAL APPROACI:! irJ
/

Le~ us tabulate the differences between the two now.

......
Al Program Cooveotiooal Program ......
1· I.
It manipulates numeric information.
It manipulates symbolic infonnation.
It uses algorithmic search method.
2· It uses heuristic search method. 2.
3. It seeks for optimal answers.
3- It seeks for satisfactory answers.
4. It involves large DB.
4. It involves large KB.
It involves rare modifications.
5. It involves frequent-modifications. 5. -
Ill INTELLIGENT AGENTS
These are software programs which work in the background to carry out specific, repetitive,
predictable task for an individual user. Alan Turing suggested that a system or agent can be said to be
intelUgent when the agent's performance cannot be distinguished from that of a·human performing
the same task. Thus, we define now. An intelligent agent is a software entity which senses in
environment and then carries out some set of operations on behalf of a user with some amount
of autonomy and to do so, it employs some knowledge or representation of the end-user's goal.
Please note that an agent is different from a program. An agent need not be a program at all. It
may be a robot or a college professor. Also note that Software agents are by definition programs but
a program must measure up to several marks to be an agent.
For example, humans, animals, autonomous mobile robots or softbots calculators are real-
world agents. Software agents live in system's O.S.
Please understand that systems are agents or not with respect to some environment. AI is about
the design and implementation of intelligent agents. The concept is shown below in Fig. 1.5.

Percepts

Environment

Action

Fig. 1.6 Role of an agent

. As shown in Fig. 1.6, an agent perceives its environment through sensors. It then acts upon that
envrronment through actuators.
For examples-
1. A human agent has eyes, ears, nose etc. as sensors while hands, legs, mouth are actuators. ,
2. So~are agenl r~ceiv:s k_eystrokes, files, packets are sensory inputs and acts on the
envrronment
. . by d1splaymg on the screen, writing
. files and sendm'g packets over a network.
Charactenstics of Agents
. Agents have some features which are listed below:
1. They are autonomous, i.e., can work on their own.
2- They are Persistent over a prolonged time period.
3· They are adaptive, i.e., adjust to changes.--
 · 11

~:::.::.::..:::....:....:..::~~- - - - - - - - - -- -
INTRODUCTION TO Al

4 · They are inobife., i.e., can be transported over ne1WoTks.

-

5. They have ability to learn.

AppUafioru of Intelligent Ag':nts (IA)
Tbere are some major applications of IA and are given below--
I . lAs are used to access and navigate information using different search engines.
2. IAs help in decision-making by the knowledge woTkeTS.
3. IA like voice-activated interface agent reduc~ the user's task of explicitly commanding
the computer.
4. IAs perf~ the time-consuming and combeTSome tasks of searching databases, doi~g
retrievals and filtering of information and sending the result-sets back to the user, m
distributed environments.
5. IAs can be used to assist manageTS to do their job. Some management-oriented tasks that
an agent can do are-- advising, alter, browse, distribute, enlist, explain, filter, guide, match,
monitor, navigate, negotiate, organize, query, report, remind, retrieve, schedule, search,
bi secure, store, suggest, summarize, reach, translate and watch.
Ill
)1l,
L SUMMARY
hll The word 'Artificial' means making, a copy of something natural and 'Intelligence' means the
abiJity to gain and apply knowledge and skills. Kurzweil, in 1990, defined AI as the art of creating
machines that perform functions that require intelligence when performed by people. But please
Rmember that humans are much more intelligent than Al system as we (humans) only program
,Oil these Al applications (called as AI programs now). Also, we found that the idea of AI originated
from a historic experiment called as Turing Test.