AI Demystified by Alex David Pratt - HTML preview

PLEASE NOTE: This is an HTML preview only and some elements such as links or page numbers may be incorrect.
Download the book in PDF, ePub, Kindle for a complete version.

DEDICATION

This book is dedicated to all true lovers of technology and every AI

tool out there.

5

TABLE OF CONTENT

FOREWORD .......................................................................................... 13

INTRODUCTION .................................................................................. 15

CHAPTER ONE ..................................................................................... 17

WHAT IS AI? ................................................................................. 17

TYPES OF AI ................................................................................. 23

Weak AI ..................................................................... 23

Strong AI ................................................................... 23

STAGES OF AI .............................................................................. 24

Ruled-Based AI.......................................................... 24

Context-Based AI ...................................................... 25

Expert System AI ....................................................... 25

Reasoning AI ............................................................. 26

DATA ............................................................................................. 26

MACHINE LEARNING ................................................................ 26

CHAPTER TWO .................................................................................... 28

WHY AI? ........................................................................................ 28

ADVANTAGES OF AI .................................................................. 28

MULTITASKING ..................................................... 28

SPEED ....................................................................... 29

DILIGENCE .............................................................. 29

AI VS MAN.................................................................................... 30

DISADVANTAGES OF AI ........................................................... 30

Misinformation .......................................................... 30

Mental Dullness ......................................................... 31

Corruption .................................................................. 31

Bias ............................................................................ 31

APPLICATIONS OF AI ................................................................ 32

BUSINESS ................................................................ 32

HEALTH ................................................................... 33

7

EDUCATION ............................................................ 33

SPORTS ..................................................................... 34

GAMING ................................................................... 34

IT................................................................................ 34

TRANSPORTATION ................................................ 35

COMMUNICATION................................................. 35

RELIGION ................................................................ 35

WORK ....................................................................... 36

SKILL SETS THAT ARE LONG LOST TO AI ........................... 36

Writing ....................................................................... 36

Art Design .................................................................. 37

Financial analysis ....................................................... 37

Language Translation ................................................ 37

Data Entry .................................................................. 37

Image and Video Editing ........................................... 38

Scheduling ................................................................. 38

AI: FRIEND OF FOE? ................................................................... 38

CHAPTER THREE ................................................................................ 41

WHEN AI? ..................................................................................... 41

PAST OF AI ................................................................................... 41

PRESENT OF AI ............................................................................ 43

ChatGPT .................................................................... 44

Gemini ....................................................................... 44

Bard ............................................................................ 44

Grok ........................................................................... 44

Imagine ...................................................................... 45

FUTURE OF AI ............................................................................. 45

KINDS OF AI ................................................................................. 46

Reactive Machines ..................................................... 46

Limited memory ai..................................................... 46

Theory of Mind AI ..................................................... 47

Self Aware AI ............................................................ 47

LEVELS OF ARTIFICIAL INTELLIGENCE............................... 47

Artificial General Intelligence ................................... 47

Artificial Super Intelligence ....................................... 48

CHAPTER FOUR .................................................................................. 51

HOW AI? ........................................................................................ 51

Automated Document Processing .............................. 51

Dynamic Fraud Detection Systems ............................ 52

Smart Supply Chain Management ............................. 52

Augmented Creativity in Content Creation ............... 52

Accelerated Product Development ............................ 52

AI in Media and Entertainment.................................. 53

Intelligent Code Generation Assistance ..................... 53

Automated Report Generation ................................... 53

Conversational Search Interfaces ............................... 53

Elevated Customer Experiences with Chatbots ......... 54

Sell AI Art Prompts ................................................... 54

Sell AI Stock Photos .................................................. 54

Sell AI Print On Demand Photos ............................... 55

AI Content Creation ................................................... 55

AI Music .................................................................... 55

AI Video .................................................................... 55

AI Coding .................................................................. 56

MUST-KNOW AI TOOLS ............................................................ 56

CONTENT CREATION ................................................................ 56

Copy.AI ..................................................................... 56

Notion.ai .................................................................... 57

Rytr ............................................................................ 57

Jasper ......................................................................... 57

DESIGN AND VISUALS .............................................................. 57

MyEdit ....................................................................... 57

Synthesia .................................................................... 57

Lumen5 ...................................................................... 57

Canva ......................................................................... 58

PRODUCTIVITY AND ORGANIZATION .................................. 58

9

Descript ...................................................................... 58

Todoist ....................................................................... 58

Evernote ..................................................................... 58

Calendly ..................................................................... 58

Bard ............................................................................ 59

MUST-HAVE AI SKILLS ............................................................. 59

1.

Prompt Engineering ........................................... 59

2.

GenAI: ............................................................... 59

3.

LLM Building: ................................................... 60

Data Science .......................................................... 60

5.

AI Innovation: ................................................... 60

BEST COURSES TO LEARN ABOUT AI ................................... 61

1.

AI for Everyone ................................................. 61

2.

Artificial Intelligence Nanodegree .................... 61

3.

Computer Science for Artificial Intelligence ..... 61

4.

Deep Learning Specialization ............................ 61

5.

Self-Driving Cars - Duckietown ........................ 61

6.

NLP Specialization ............................................ 61

7.

Artificial Intelligence ......................................... 61

ADDITIONAL PLATFORMS TO LEARN ABOUT AI .............. 61

1.

Google AI (ai.google/education): ...................... 61

2.

IBM Skills (ibm.com/training): ......................... 62

3.

TensorFlow Tutorials (tensorflow.org/tutorials):62

4.

PyTorch Tutorials (pytorch.org/tutorials):......... 62

5.

Udacity (udacity.com): ...................................... 62

6.

Codecademy (codecademy.com): ...................... 62

7.

Kaggle (kaggle.com): ........................................ 62

8.

Fast.ai (fast.ai): .................................................. 62

9.

LinkedIn Learning (linkedin.com/learning): ..... 62

10.

DataCamp (datacamp.com): .......................... 62

BONUS ........................................................................................... 63

Prompt Engineering Basics: .............................. 63

ChatGPT Prompts Mastery: ............................... 63

Intro to Generative AI: ...................................... 63

AI Introduction by Harvard: .............................. 63

Microsoft GenAI Basics: ................................... 63

Prompt Engineering Pro: ................................... 63

Google's Ethical AI: .......................................... 63

Harvard Machine Learning: ............................... 63

LangChain App Developer: ............................... 63

Bing Chat Applications: .................................... 63

Generative AI by Microsoft: .............................. 64

Amazon's AI Strategy: ....................................... 64

GenAI for Everyone: ......................................... 64

AWS GenAI Foundation: .................................. 64

OpenCV Bootcamp: .......................................... 64

Tensorflow Bootcamp: ...................................... 64

GLOSSARY ........................................................................................... 64

REFERENCES ....................................................................................... 66

ABOUT THE AUTHOR ........................................................................ 69

11

FOREWORD

This book is more than just a technical manual. It's an

invitation to a grand ball, where robots waltz with algorithms, and

the dance floor shimmers with the possibilities of a future sculpted

by AI. You'll witness mundane tasks blossom into graceful

automatons, freeing us to paint on the canvas of human potential

with bolder strokes. Imagine surgeons, empowered by AI insights

invisible to the human eye, performing miracles of healing. Or

artists collaborating with digital muses to birth masterpieces never

before dreamed of.

But, as with any potent potion, AI demands responsible

consumption. This book unflinchingly confronts the shadows that dance at the edge of progress. It exposes the dark alleys of deep fakes, the slippery slopes of automation, and the ever-present question: what happens when the spark of intelligence becomes an

uncontrollable inferno?

"AI Demystified" is a clarion call, not a lullaby. It's a reminder that the future of AI is not written in lines of code, but in the choices we make today. With each mindful step, each responsible decision,

we weave a tapestry of progress, ensuring that AI is not merely a force to witness, but a co-pilot in the grand adventure of humanity.

So, turn the page, dear reader, and let Alex David Pratt guide you

through the maze of misconceptions. Embrace the spark, for within

it lies the power to illuminate, to empower, and to shape a future as boundless as the human imagination itself.

Bard.

13

INTRODUCTION

" AI is here to complete and not to compete with human intelligence"

– Alex David Pratt

Artificial Intelligence or AI in short is a buzzword that has clocked a staggering search volume of over 600,000.

Artificial Intelligence or AI is a hot trend in the digital space, majorly because of ChatGPT and other mind-blowing A.I.

algorithms that have gone viral recently. ChatGPT is a natural language processing software that generates sensible output based

on the prompt you provide it.

As fascinating as ChatGPT is, you'd be amazed to discover that

it is just an inferior iteration of an even more mind-boggling phenomenon. Beyond natural language processing, AI retains the future of medicine, law and even religion.

Why is such a humungous amount of people interested in a

topic? Well, quite simply because 400 out of the Fortune 500

companies utilize this fascinating technology to handle a spate of their tasks.

The global AI market is expected to reach a stupefying 1.81

trillion by 2030 and is predicted to increase by a CAGR of 38.1%

The exceptional rate at which it is advancing makes it a subject

that everyone needs to be enlightened about if they are to remain relevant and thrive in this fast-paced digital age.

A good number of people have heard about AI but hold many

negative notions about what it is. To some AI is a robot or a single

company’s product, but what is it? In this short but riveting piece, I will be defining all the essential terms you need to be informed about as it regards Artificial Intelligence.

15

In this comprehensive piece, we will be journeying from the very birth of AI in the Dartmouth College campus to its incredible

future. We will be looking at all that AI could possibly become with time and how that will impact human life.

After reading this book, you will have gained a deeper

understanding of what AI is and what it holds in store for mankind.

Apart from granting you an understanding of the many benefits and advantages of AI, I will also be showing you how you can utilize it to your advantage.

It is not enough for you to just be aware of the concept of Artificial Intelligence and how it will significantly improve the human race, beyond this; you need to know how exactly it can be

of help to you.

As you read this book, do understand that not all of the content

reflects my opinion nor do I support some of the negative

implications AI may have on humanity.

My interest is to simply enlighten the audience on the

implications both positive and negative so that prudent decisions can be made to regulate this intriguing but dangerous innovation.

CHAPTER ONE

WHAT IS AI?

In the end, intelligence will become so much more artificial than it is biological.” – Alex David Pratt

AI as a compound word may seem a bit hard to define. Of

course, it entails the study of intelligent agents (machines capable

of perceiving their environment and taking actions to maximize the

chances they have to achieve their goals); it is so broad that it

probes into the realms of philosophy, psychology, sociology,

biology, neuroscience and even spirituality. The abbreviation AI

stands for " Artificial Intelligence" and to some extent the name is self-explanatory but I will expatiate on what it entails. Firstly, I’d like to take a keen look at the word ‘ Intelligence’, for though it is the second word, it is the keyword in the abbreviation.

Intelligence could simply refer to the ability to think or reason

that is akin to humans and this also encompasses the capacity to

learn, solve problems, comprehend and perceive relationships.

Intelligence also means to be able to grasp information, relate it

with other information, understand it and manipulate it in different

ways. To further understand what intelligence is, we need to study

mankind, who is the epitome of it.

Humans are the most intelligent beings in the universe (this is

as much as we know for now) and there are several reasons why

we deem them so; a few of them include their ability to reason

(think critically and analytically to make an informed decision).

Our learning abilities and advanced problem-solving skills as

17

ALEX DAVID PRATT

humans are other cognitive powers that make us the apex creatures

in existence.

Even our capability to understand ourselves, relate with others

and interact with our surroundings through our senses are results of

intelligence. Intelligence is ability; it is the potential to do a wide variety of things; if someone can easily tell apart different musical notes, we say he is musically intelligent because He has knowledge

and understanding of that realm.

Human intelligence is multifaceted and multi-dimensional

different categories of intelligence are attributed to men. A few of

them include are: linguistic intelligence (understanding

information in a spoken or written language and giving meaningful

responses, creative intelligence (bearing the ability to come up

with new ideas), and spatial intelligence (being able to perceive

space, comprehend distance and location). The list goes on and on

about the many unique mental abilities that men possess.

Beyond the mere capability of thought, intelligence also entails

being able to think reasonably and act rationally. Being truly

intelligent means being able to think right and act right as a result it is not limited to you just thinking alone.

Having gained an understanding of what intelligence is, let’s

move to defining the word “Artificial”. Saying something is

artificial is saying that it is unoriginal. It also denotes that it is unreal and contrived. Juxtaposing the words artificial and

intelligence makes it plain that AI is intelligence that is unoriginal but that doesn’t mean it is unreal.

Artificial Intelligence can be defined in many ways but I’d like

to define it as the simulation of abilities typically attributed to humans, by machines. It can be explained in terms of machines mimicking human cognitive abilities, rational thinking or action. It

is the imitation of human intellect or behaviour by machines. To

AI DEMYSTIFIED

simulate simply means to model or imitate something and in this case it is when machines reproduce abilities (mostly cognitive) that

were originally exclusively human. Apart from mimicking

intellectual capabilities AI can also be considered the as the imitation of perceptual, emotional and social qualities that are predominantly ascribed to humans alone. Artificial intelligence is termed so for a reason, saying something is artificial already means

it is not original, it proposes that it is a replication of something and that is exactly what AI is. It is when devices display intellect that only humans are supposed to be able to do.

For instance, I spoke about linguistic intelligence which is what

powers most of our communication, i.e., our ability to grasp

written or spoken language. This can easily be replicated by

machines through Natural Language Processing and other human

cognitive abilities (our ability to assimilate and manipulate new

information) can also be contrived by machines through

knowledge representation and automated reasoning.

AI can possess a goal and discover ways to go about

accomplishing it just as any human would, only that, the process it

would take to reach a goal will be purely based on data analysis

and mathematical methods rather than the common sense or

natural practical judgment a human would employ to achieving it.

AI only simulates it doesn’t have intelligence as a human does.

AI can simulate certain kinds of intelligence quite remarkably

but there are dimensions of intelligence it cannot even attempt to

replicate. Creativity for example is one of the unique intelligences

of man that computers are yet to demonstrate because creativity

deals with coming up with new patterns of thoughts. You may

argue that ChatGPT or other generative AI seem to be capable of

coming up with songs, text and images when prompted but the

19

ALEX DAVID PRATT

truth of the matter is that, generative AI only studies patterns from the data you provide it and spawns mathematical variations of it, so

in reality AI doesn’t create (comes up with something completely

new).

Another distinct human trait that machines will never be able to

possess is emotional intelligence which stems from interpersonal

and intrapersonal knowledge. For computers to generate emotions,

they must be self-aware and must have personal goals, interests

and desires. This is quite impossible today because scientists are

still grappling with comprehending what the consciousness (seat of

personality) is much less simulating it. The more interpersonal an

intelligence is, the harder it is for AI to imitate. This reminds me of the imitation game, which I will talk about later on.

Computers are far away from becoming sentient (knowing and

becoming conscious themselves) but they can in a sense know

others. Alan Turing, a notable name in the AI arena conducted

several tests which assess computer intelligence depending on

whether a human can tell a computer apart from a person. The first

of these two tests is

Have you ever supplied a computer with information and it

responded to you in a manner so accurate that you felt like it was

human? Yes, this may have happened to you before, especially if

you interacted with many of the AI Chatbots available online.

However, the mere fact that they provide intelligible responses

doesn’t mean they possess any level of intellect because they may

have merely been programmed to recognize a particular sentence

syntax and prove a specific response to it, this does not mean that

the computer grasps the message being related to it. Although

these sensible responses make it look like they do understand

human thoughts and feelings, the opposite is true because

computers simply cross-reference terms in a lookup table when

AI DEMYSTIFIED

certain keywords are used and act on the algorithm contained in

the table to give a suitable response in a very mathematical rather

than interpersonal fashion.

To prove this point, an AI thought leader, conducts an

experiment where a person is locked in a room and provided with a

library of rules and lookup tables in Chinese. If the person inside

receives a note in Chinese from someone outside; he can easily

search for the characters on the note from the dictionaries in the

room and decipher what the person is trying to communicate but

that does not mean he understands the Chinese language.

Merely knowing something doesn't mean you understand

which equates to intelligence. True intelligence is in understanding

and ability. This poses a question, is an AI computer able to think,

grasp and understand the concepts that are fed to it? All AI has is a predefined set of rules that tells it how to respond in every case;

coupled with the ability to learn it can now also learn from various

instances just like a human would. As children growing up, we

learn from our mistakes, and successes. We do good and we are

rewarded then we automatically know that the actions we took are

to be repeated shortly if we want to have such rewards again; we

do something bad and get a beating; that helps us to know that

such is not something to do again.

For AI to compete with human intelligence it needs to replicate

the same processes of human intelligence, however, that’s not the

case. In contrast, AI demonstrates intelligence yet uses other

means to reach an end and that is why it is not intelligent to the

human level.

Although certain machine intelligences can parallel certain

human intelligences how they function is quite antithetical.

Computer vision can be an easy mechanical rival to visual

21

ALEX DAVID PRATT

intelligence or robotics an alternative to human psychomotor

intelligence but the dynamics behind a human’s kinaesthetic

prowess entails recondite abilities such as reflexes that are beyond

the reach of AI imitation.

Humans identify a goal; they want to achieve something, then

they go all out into seeking ways that are attainable and then learn

from the failure or success of attaining that goal to either make

another approach or otherwise improve upon the previous ones.

These processes can be simulated but they are not an exact

copy. The major difference between AI and human intelligence is

in the processes they use to function. Artificial Intelligence uses

more logical and mathematical modes of operation, learning

patterns and utilizing that knowledge to extrapolate.

Human intelligence on the other hand is far more sophisticated

especially because of the myriad neural networks that it functions

upon. The way the human brain works is so complex to the point

that opportunities and even novel ideas or ways things can work

are just mere connections we realize between disparate objects.

Psychologists and neuroscientists are still yet to fully understand

how the brain works much less replicate it in a computer system.

Contrary to popular opinion, AI is not a robot! When some

people think about AI, they just envision a Terminator-like entity,

or others think of fictional characters such as iRobot, etc.

Although, robots could be powered by AI; that is not really what

they are. Robotics mainly AI to simulate human proprioception;

nonetheless, these two disciplines are quite distinguishable.

AI is not a single company’s product! Some people have

limited their definition of what AI is to Apple’s Siri, Amazon’s

Alexa or OpenAI’s ChatGPT, now while these are all AI

technologies, they are merely one wave of AI and not all that it is.

AI DEMYSTIFIED

Siri, Alexa, and Cortana are virtual assistants that fall under the

category of AI. They are customized search engines and digital

assistants that help users easily find certain pieces of information

they are looking for and further help to perform tasks as simple as

turning on a light bulb to tasks as sophisticated as finding the

nearest restaurant that sells tacos.

AI can be categorized in different ways, there are different

types, stages, kinds and levels of AI. I’ll talk about the types and

stages of AI and then talk about the kinds and levels of AI in

subsequent chapters.

TYPES OF AI

WEAK AI

Narrow AI or weak AI is specially trained to perform a

particular task and solve a specific problem. It may also be termed

an expert system, that gains mastery in a particular domain. As the

name implies, the task that this AI is designed to perform is

narrowed down to just a single thing or two. Siri, Alexa and

Cortana are good examples of this, they’re good at just one thing

and that is all they focus on, helping you find relevant information.

These iterations of AI applications were built on learning models

that required human intervention to process new information or

carry out any task that was beyond their initial training.

STRONG AI

Strong AI is capable of accomplishing a wide range of tasks

and knows a variety of fields. Strong AI is still under development

23

ALEX DAVID PRATT

and we may not fully see it anytime soon, it can also be termed

Artificial General Intelligence and it utilises the knowledge it’s

attained from previous learning experiences to accomplish new

tasks in contexts it hasn’t encountered before. Weak AI cannot

accomplish this without receiving further training in the underlying

models through human intervention.

Strong AI will prove very useful in robotics since it greatly

demands the ability to make decisions beyond its training as it

explores the real world that is laden with miscellaneous

environments. Kinaesthetic intelligence is one of the many

intelligences advanced robotics will require because it has many

moving parts and would need to know how to utilize them to

perform different tasks but significant advancements are being

made in this area, in fact, most recently an AI robot beat humans in

a physical skills game.

STAGES OF AI

RULED-BASED AI

AI started in a rule-based form; i.e., it needed to be instructed

on what to do for it to function effectively. This AI is a knowledge-

based system; it makes decisions based on predefined rules and

doesn’t possess the ability to learn or adapt to unexpected

situations and is therefore limited in its functionality. This stage of AI is readily found in technologies such as thermostats and alarm

clocks; an alarm clock chimes at a particular time because it has

been programmed to do so and a thermostat may turn on the air

conditioner when the temperature is at a particular degree. Radios

and microwaves also utilize rule-based AI technology.

AI DEMYSTIFIED

CONTEXT-BASED AI

This stage of AI doesn’t just accept and process input it takes

note of factors such as background data, environment, and real-

time user behaviour to make decisions. Siri and Alexa fall in this

category they can make wise suggestions based on your personal

preferences that they’ve taken note of over time. They can

remember every encounter with you and store it for future

purposes. Google Discover uses context-based AI to recommend

sites to you based on your browsing history. Context-based AI

plays a major role in helping you tailor your experiences on the

internet by only appealing to your interests.

EXPERT SYSTEM AI

This stage of AI deals with AI systems that are trained and

taught to master specific fields and domains to the point that they

can act in the stead of an expert in that field. For example, an

expert system in the medical field may be able to prescribe drugs

to you if you can explicitly describe your symptoms.

They’re able to consume vast volumes of data and provide

valuable insights that would take humans longer hours to study and

uncover. AlphaGo is an example of such a system, only that it was

created to master the board game Go.

The financial world maximizes such systems to monitor

markets and predict stock movements with unbelievable accuracy.

Real-time translation is another amazing ability this AI wields.

Samsung's S24(the AI Phone) features this fascinating technology.

25

ALEX DAVID PRATT

REASONING AI

This stage of AI can simulate the composite processes of

humans though. They can analyse information draw connections,

spot errors and do some complex things that are only possible with

human-level cognitive powers. They can even bring unseen or

unobvious possibilities to light. ChatGPT can be said to be in this

stage, it is a Large Language Model (LLM) that’s built on training

data from myriad websites. Autonomous driving cars are also said

to be in this stage of AI.

DATA

AI is only as powerful as the amount of data that is used to

train it. Just like how food nourishes and replenishes our bodies

causing it to grow and develop, data is food to AI. In the same way

that we do not eat raw food but have to prepare and process it well

before it is edible, data has to be manicured before it is usable. Not all data is useful because not all data is truthful; hence data has to be trimmed and put into the right shape and mode for it to be

useful. As I explained earlier, AI cannot create, so for it to give an output, it must first be supplied an input. If data is what AI feeds

on to learn you may be wondering what process AI use to learn, I

treat this right below.

MACHINE LEARNING

Machine Learning (ML) is the autonomous learning process of

A. I. machines, refers to the process of them learning from data on

their own accord. Usually, machines would need to be told what to

do and rely on external input of data for them to enhance their

AI DEMYSTIFIED

performance; such isn’t the case with ML where the machines can

improve their performance by recognizing patterns and making

predictions. ML enables AI to gain intelligence from data and learn

from each experience to detect new patterns that it could use to

easily adapt to similar circumstances.

Machine learning has three major paradigms: there’s

supervised learning, unsupervised learning and reinforcement

learning. The first as the name implies entails where an input

object and desired output value are used to train a model in a

monitored environment. Unsupervised learning is similar to

supervised learning only that it is within an environment where the

output values aren’t provided, but it can infer some things by

clustering. Reinforcement learning enables an AI model to learn

through reinforcements (rewards or punishments).

Deep Learning is a kind of ML that copies the way your brain

works. Just as countless neural networks in your brain are

interconnected deep learning seeks to mimic this intricate

architecture. AI based on deep learning will be able to make more

accurate judgements and make realizations that would be much

harder to discover through RL alone.

Deep learning empowers AI applications to learn how to

execute jobs that will require human intelligence; this enables them

to learn just like humans do to easily make decisions in new

environments beyond their initial training.

Computers can generate content and automate tasks because of

advancements in deep learning, however, certain aspects of this

form of AI remain theoretical and under development. DeepMind

is a Google AI research company that is doing an impressive job of

exploring deep learning.

27

ALEX DAVID PRATT