Essentials of Generative AI

Getting Started with GenAI

Engage and Think

Why do you think generative AI is causing such a

disruption across various industries when AI
technology has been around for years?
 Learning Objectives

By the end of this lesson, you will be able to:

Identify the applications of GenAI in solving real-world problems

Create prompts to communicate effectively with Gen AI tools for
desired outcomes
Apply GenAI tools to create personalized documents such as resumes
or conduct research
Use powerful GPTs, such as Write For Me, Canva AI, DesignerGPT,
Consensus, and Universal Primer, to explore diverse Gen AI capabilities
for various tasks
Introduction to Generative AI
 Evolution from AI to Generative AI

This timeline visually represents significant milestones in the field of artificial intelligence.

1950 ELIZA 1980 2014 2018 GPT

LSTM Transformer
Chatbot series

GRU and
Turing Neural Attention LLM 2020 -
1966 network 1997 2017
test mechanism Present
 Generative AI

Generative artificial intelligence (GenAI) is a category of AI systems that focus on generating new
content, such as images, text, music, or other forms of creative output.

It is a subset of artificial intelligence that leverages machine learning and deep learning techniques
to generate data.
Objectives of Generative AI

It is designed to independently generate new content by

deriving its structure from recognized patterns.

It assists with complex problem-solving by generating

numerous potential solutions, thus offering users a wide
array of options.

It boosts human creativity by producing unique outcomes,

enabling users to discover new ideas and possibilities.
 Generative AI and LLM

Large Language Models (LLMs) provide a foundation for generative AI by offering powerful,
pretrained models capable of understanding and generating human-like text across various
tasks and domains.

Generative AI can serve as building blocks for more advanced models, enabling the development of
innovative AI systems that create content ranging from natural language text to images and music.
 Large Language Models(LLM)

LLMs are state-of-the-art AI models designed to comprehend and generate human language and
perform natural language processing tasks like Translation, Classification, and so on.

Refers to the significant size and complexity of these models, which

Large
contain hundreds of millions or even billions of parameters

Denotes their primary function, which is to understand and generate

Language
human language

Describes them as mathematical representations that capture the

Model
patterns and structure of language data
 Widely Used LLMs

Model Developer Key features

It helps generate content, text summarization, sentiment analysis, data

extraction, and language translations. The GPT 4 upgraded version
GPT 3.5/4 OpenAI
bolsters multimodal capabilities, including the power to process
different formats other than texts, like images and videos.

Gemini Google It aids in complex problem-solving and code generation.

It enables natural language understanding, content creation, and

Llama Meta AI
language translation.
Technology It provides language translation, common-sense reasoning, and
Falcon Innovation
sentiment analysis.
Institute(TII)
It supports open-ended dialogues, coding tasks, and data analysis of
Claude Anthropic
visual inputs like charts and graphs.
 GenAI Tools

Top GenAI tools transforming various industries:

Demo: Resume and Cover Letter Creation Using ChatGPT

Duration: 10 minutes

Learn how to create a resume and a cover letter using ChatGPT.

Imagine you have a resume ready, but you're unsure if it matches the job description you're applying
for. Additionally, you're pressed for time to update it according to the job requirements.

In this scenario, how can you quickly and efficiently update your resume to align with the job
description and save time?

Join us in a demonstration where we will use ChatGPT to help you quickly and effectively tailor your
resume to the job description.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
Quick Check

As an intern at a tech company, you've been tasked

with choosing an LLM for a project that requires
generating highly accurate and factual content for
technical documentation. Which of the following LLMs
would be best suited for this purpose?

A. GPT-4
B. Gemini
C. Falcon
D. Claude
Understanding Prompts
 What Is a Prompt?

A prompt is a natural language text that instructs generative AI models to perform a specific task.

Here are some key reasons why prompts are essential:

A precise prompt can lead to a more

focused and useful output.

A well-crafted prompt guides the GenAI

to understand the context and the
specific requirements of the task at hand.

The prompt serves as a bridge between

the user's intent and the AI's execution.
 How Prompts Work?

When you give a prompt to a language model, it generates a response based on the patterns and
information it has learned during its training.

Here's a simple example:

• Here, the prompt is a straightforward question, and

the language model provides an answer based on its
training data.
• Prompts can also be more complex and used for
various tasks such as text completion, translation,
summarization, and even creative writing.
• The effectiveness of the response often depends on
how well the prompt is crafted.
 Important Components of a Prompt

Four important components that make up a great prompt are as follows:

Task Instructions Context Input

Background
Specific action Rules to follow Images, text, and
information or
or process to achieve the dataset
situation that
desired outcome
surrounds a task
 Important Components of a Prompt: Example

Consider the below example to understand the important components of a prompt better:

Task: Create a brief narrative.

Instructions: The story must involve a wizard's quest and include a surprise ending.

Context: The narrative is set in a magical forest, with the protagonist on a quest.

Input: Use the provided lore and descriptions of creatures within the forest.
Quick Check

Which component should be included in a prompt for

a language model to translate a technical document
from English to French?
A. Specify the source and target languages in the
instructions.
B. Use a dataset unrelated to the technical field of the
document.
C. Exclude any background information about the
document's content.
D. Include technical jargon without explanations.
Hands-on with ChatGPT
 Day-to-Day Tasks ChatGPT Can Do

ChatGPT is adept at handling a multitude of complex tasks beyond mere conversation.

Here are a few notable examples:

LinkedIn
Language profile Sentiment
translation creation analysis
Demo: Multilingual Book Translation Using ChatGPT

Duration: 10 minutes

Demonstrate how to translate user reviews in various languages using ChatGPT.

Imagine you operate an online bookstore that is popular worldwide. You offer books in multiple
languages to cater to a diverse, global audience. To expand your reach and provide readers with books
in their preferred languages, you use ChatGPT for real-time book translation.

To explore this further, let’s consider a demonstration in which we'll utilize ChatGPT to facilitate the
translation process and demonstrate its effectiveness in handling language barriers.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
 Demo: LinkedIn Profile Creation Using ChatGPT

Duration: 10 minutes

Demonstrate how to create an impressive LinkedIn profile using ChatGPT.

Imagine you're preparing to create your LinkedIn profile, but you're aiming for something more exciting
by creating a personal brand.

Let's check out this demo, where we'll use the ChatGPT prompt to help make your LinkedIn profile
awesome and stand out from the crowd!

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
Demo: Sentiment Analysis of User Reviews Using ChatGPT

Duration: 10 minutes
In this demo, we’re showcasing sentiment analysis with ChatGPT to distinguish between praises,
complaints, and many other types of feedback.

Have you ever scrolled through tons of online reviews, wondering which ones are positive and which
are negative? This can be time-consuming.

ChatGPT can instantly analyze the reviews and provide the percentage of positive and negative
feedback.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
ChatGPT: Multimodal Capabilities
 ChatGPT: Multimodal Capabilities

The advanced version of the ChatGPT (ChatGPT – 4) model has multimodal capabilities that go
beyond just text generation.

Text-to-speech/
Text/Document Image
Speech-to-text

Note: ChatGPT has the capability to convert text-to-speech and speech-to-text in the mobile version.
 ChatGPT: Multimodal Capabilities

• Integration of multiple modalities: Combines text, images, videos, audio, and more to
understand and generate responses
• Enhanced context awareness: Utilizes visual and auditory cues to provide more
comprehensive and accurate answers
• Dynamic and interactive responses: Enable a more natural and human-like conversation
experience
• Image captioning: Describes images using natural language
• Video analysis: Extracts information from video content and responds accordingly
• Virtual assistants: Offer image-based responses for enhanced assistance
Demo: Exploring Multimodal Capabilities of ChatGPT

Duration: 10 minutes

Demonstrate how ChatGPT works beyond text.

In this demo, our objective is to delve into the realm of multimodal tasks, where ChatGPT will be
equipped to handle both visual and textual inputs concurrently.

Through this endeavor, we aim to showcase the potential of ChatGPT in addressing complex
multimodal tasks and advancing the capabilities of GenAI systems.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
Demo: Customer Feedback Analysis Using ChatGPT-4

Duration: 10 minutes
Demonstrate how to use ChatGPT to categorize customer feedback into positive, negative, constructive,
or neutral sentiments and visualize the results.

As a data analyst at a large e-commerce company, your task is to analyze the customer feedback
collected after a major sales event. The objective is to identify areas of success and opportunities for
improvement, thereby aiding the company in enhancing its future sales strategies.

To address this, let's explore a demo where we'll experiment with using ChatGPT prompts to tackle this
issue.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
Explore Powerful GPTs
 Comparison between ChatGPT 3.5,4, and 4o

Here is a comparison highlighting key differences:

Feature ChatGPT-3.5 ChatGPT-4 ChatGPT-4-turbo (4o)

Model size Medium Large Large

Context length Up to 4,096 tokens Up to 8,192 tokens Up to 128,000 tokens

Speed Fast Slower than 3.5 Faster than 4

General conversation, Complex tasks, Complex tasks,

Use cases
basic tasks nuanced understanding faster responses

Optimizations Basic Advanced Turbocharged for efficiency

Multimodal
No Yes Yes
capabilities
 ChatGPT Comparison Based on a Use Case

Scenario: A researcher needs to summarize a 20-page research paper to identify key findings,
methodologies, and conclusions.

Responses ChatGPT-3.5 ChatGPT-4 ChatGPT-4-turbo (4o)

Produces a detailed and

Provides a general
Summary accurate summary, capturing Matches the detail and accuracy
summary but may miss
quality nuanced details and of ChatGPT-4
some nuanced details
methodologies

Speed Fast response Slower compared to 3.5 Faster than ChatGPT-4

Handles very long context

May struggle with longer Handles longer context better,
Context efficiently, making it ideal for
context, requiring multiple providing a more cohesive
handling summarizing lengthy
prompts summary
documents in one go
 Comparison Based on a Prompt

Here are the responses from different ChatGPT models based on a single prompt.
Prompt: Explain polymorphism to me in a line

ChatGPT-3.5

ChatGPT-4

ChatGPT-4-
turbo (4o)
 Explore GPTs: Categories of GPTs

The explore GPTs are custom GPTs accessible only with the GPT-4 version.
The Explore GPTs can be categorized in different ways, as shown
below:

Writing

Education
Productivity

Research and
Programming
analysis
 Writing GPTs: Write For Me

This GPT is focused on writing, such as content generation, writing assistance, or editing.

Write For Me is a personal writing

assistant. It crafts tailored content,
focusing on needs, audience, and SEO,
ensuring ideas are transformed into
engaging written material.
Demo: The use of the Write For Me GPT for creating
Marketing Content

Duration: 10 minutes

In this demo, you will be able to illustrate the use of the Write For Me GPT for creating marketing
content for an eco-friendly reusable water bottle, aiming to connect with environmentally conscious
consumers.

The purpose is to demonstrate the application of GPT technology in creating engaging content that
resonates with the target audience's environmental values.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
 Productivity GPTs: Canva AI

This helps with increasing efficiency in tasks or workflows.

Canva AI makes designing easy by

creating designs from your descriptions.
Perfect for beginners and pros, it helps
you make social media posts and
presentations and turns text into images.
Demo: Creative Branding Makeover for Coffee Shops with
Canva GPT

Duration: 10 minutes

In this demo, you will be able to showcase Canva GPT's role in boosting productivity by designing
appealing content for a coffee shop's new product, Sunrise Brew.

This demonstrates how GPT technology can streamline the design process and enhance marketing
productivity.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
 Programming GPTs: DesignerGPT

This helps with coding, such as automatic code generation, code reviews, or even educational tools for
learning programming.

DesignerGPT is a specialized AI designed

to instantly create and host beautiful,
responsive HTML web pages tailored to
user requests, streamlining web design
with precision and efficiency.
Demo: Creating a Website Design Using DesignerGPT

Duration: 10 minutes

In this demo, you will be able to use DesignerGPT to generate a visually appealing website design or
product landing page for an online store, focusing on Programming GPT.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
 Research and Analysis GPTs: Consensus

This aids with research, data analysis, or information synthesis.

Consensus is a scientific research

assistant leveraging AI to synthesize
insights from academic papers, providing
clear, concise summaries and evidence-
based responses to your inquiries.
Demo: Streamlining Literature Surveys on LLM Impact with
Consensus GPT

Duration: 10 minutes

In this demo, you will be able to use the Consensus GPT for research and analysis to conduct a
literature survey on the impact of Large Language Models (LLMs) on natural language understanding
(NLU) tasks.

The goal is to gather recent studies on LLM advancements, particularly in question answering, text
summarization, and sentiment analysis, and analyze key trends and challenges while highlighting the
role of fine-tuning and transfer learning.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
 Education GPTs: Universal Primer

This encompasses GPT for teaching, tutoring, or creating educational content.

Universal Primer is the fastest way to

learn everything about anything.
Demo: Enhancing Learning with the Universal Primer GPT

Duration: 10 minutes

In this demo, you will be able to use the Universal Primer GPT to enhance educational content on
chemical equilibrium for high school students.

The main goal is to demonstrate the Education GPT tool's ability to improve learning materials and help
students understand complex concepts.

Note: Please download the demo documents from the LMS Reference Material Section for a
detailed explanation.
 Key Takeaways

GenAI is a subset of artificial intelligence that leverages machine

learning and deep learning techniques to generate data.
LLMs can be classified in several ways, depending on their functions
and users.
Prompts guide the GenAI to understand the context and specific
requirements of the task.
GenAI tools perform tasks of various complexities, such as creating
resumes, LinkedIn profiles, email classification, sentiment analysis,
and other day-to-day tasks.
The multimodal capabilities of ChatGPT can simultaneously process
and generate varied data types, such as text and images.
 Additional Resources

• Top Generative AI Tools: Boost Your Creativity

Q&A