Everything AI combined (by Shaikha)

What was the leading search engine before AI?

Google

What are AI search engines?

A tool that uses AI to improve your search experience

AI search engine features

• Understand your intent - even when phrased poorly or incomplete

• Provide comprehensive and relevant answers

• Personalized search results based on interests and past searches

• Features like imagine generation and conversational search

General purpose AI search engines:

• Bing

• ChatPDF

• Gemini/Google Bard

Research and academic focused AI search engines

• Perplexity AI

• Writefull

Academic and discovery-based search engines

Aisha

Bing

Uses LLM like GPT-4, visual models and image generation models that can:

• Search for specifics

• Provide resource links

• Let you ask follow-up questions with AI chatbot interaction

ChatPDF

An innovative AI-powered tool that allows users to have natural language conversations with PDF documents.

Uses NLP to read, comprehend and extract information with quick responses to questions and queries.

Gemini

Google’s flashship AI model

Designed for generating text, images and other content like Chatgpt

Perplexity

A search engine that uses AI to provide users with more than just basic links to websites.

• Can answer questions from simple facts to complex queries using up-to-date sources.

• Allows users to generate code, summarize articles and write emails.

• Users can organize search results in “collections”

• Allows interaction with data like asking questions about files

Aisha

Created by the ZU library as quick and efficient research support using the library resources and LLM like:

ChatGPT, Gemini, LLama 3, Claude and Mistral

and she can provide you with sources.

CRAAP test

The CRAAP test provides you with criteria to
evaluate sources for your research.

CRAAP: Currency

The timeliness of the information.

When was the information published?

Is it recent?

Is it too out of date, depending on the topic?

CRAAP: Relevance

The depth and importance of the information.

Is the material useful for your research?

Can it be used as background information or help you focus on something specific?

CRAAP: Authority

The source of the information.

Does the author have expertise on the topic?

What is their educational background?

What is their experience on the topic?

CRAAP: Accuracy

The reliability of the information.

Do you see evidence of bias?

Is the material viewed from more than one point of view?

Does it include facts supported with evidence?

Are there citations for the information presented?

CRAAP: Purpose

Why was this written?

What was the intent of the author?

Who is the intended audience?

How do we summarize manually?

1. Read the text

2. Underline important words

3. Identify main ideas

4. Write summary using TAMS

AI summarizing tools

• Quillbot

• Scholarcy

• Hypotenuse AI

Quillbot

• Top AI summarization tool that gives you creative control

• Can integrate into Word or Google Docs

• Can check synonyms, offer concise summaries and check for plagiarism

• Useful for quickly understanding long articles

Scholarcy

Specialized tool for summarizing and generating flashcards for research papers, books and academic texts.

• Uses both extractive and abstractive techniques to make summaries

• Generates flashcards LINKED to sources

Hypotenuse AI

• Can generate summaries from text inputs

• Able to handle various input options from copy-pasting to PDFs or YouTube links

• Lets you edit , add or delete

• Multiple output options like bullet points and paragraphs

AI for Writing Assistance: Grammarly

• Checks for grammar, punctuation and spelling and style errors

• Suggestions for clarity, engagement and delivery tone 

• Great for general academic writing

AI for Writing Assistance: Trinka

• Built for academic and technical writing

• Checks for subject-specific language conventions, formal tone and complex grammar issues

AI for Paraphrasing and Originality: Wordtune

• Rephrases sentences to be more formal, concise or clear

• Refines your writing style for better readability

• Rewrites sentences instantly while preserving original intent

Grammarly

• Real-time grammar and spelling checks

• Tone detection to align with the intended emotion

• Plagiarism detection tool

• Language enhancement suggestions

• Personal writing statistics and insights

ProWritingAid

• More in-depth than Grammarly

• Over 25 detailed reports

• In-depth analysis covering style, readability and more

• Helpful explanations for each suggestion

Ginger

• Grammar checking, sentence rephrasing and translation

• Many platforms and browsers

Trinka

• Specially designed for academic writing

• Features like: Word count reduction, unbiased language

• Advanced grammar checks

• Publication readiness 

Wordtune

• Rewording tool

• Rewrite or rephrase email content, articles and real time messages

• Summarize videos and text

• Check grammar and spelling

Linguix

• Grammar correction

• Ai assistance across sales, marketing and customer service

• Analyzes text quality with content score

All tools mentioned

1. Grammarly

2. ProWritingAid

3. Ginger

4. Trinka

5. Wordtune

6. Linguix

What is a prompt?

A prompt is a specific instruction or suggestion given to guide someone or something in creating or doing something.

What is prompt engineering?

Prompt Engineering is the process of creating and improving instructions (prompts) to communicate with AI systems effectively.

The 2 Types of prompts

Open-ended

Closed-ended

Open-ended prompts

Prompts that encourage detailed and expansive answers

Example: “Tell me about your day.”

Closed-ended prompts

Prompts that seek specific, concise answers

“Did you enjoy your day?” (Yes/No)

“Is it raining?” (Yes/No)

What are the components of a good prompt?

• Specificity

• Clarity

• Relevance

• Context

What are the 3 challenges in prompt engineering?

1. Ambiguity: when a prompt can be interpreted in multiple ways

2. Vagueness: lack of detail or specificity in the prompt

3. Biases: prompts that reflect personal, cultural or societal biases

Natural Language Processing (NLP)

A field of AI that focuses on the interaction between computers and humans through natural language.

Machine Learning (ML)

A subset of AI focused on building systems that learn from data.

AI Model

A computational framework or system trained to perform tasks that typically require human intelligence, such as understanding language or recognizing patterns.

Step 1 of NLP breaking down prompts

Tokenization

Tokenization is like taking a sentence and splitting it into individual words or smaller units called "tokens."

Step 2 of NLP breaking down prompts

Parsing - understanding structure

Parsing is like diagramming a sentence in grammar class. The AI analyzes how these tokens (words) relate to each other, identifying parts of speech and the overall structure of the sentence.

Step 3 of NLP breaking down prompts

Grasping the meaning/context

The AI goes beyond the literal meaning of the words and considers the broader context of the prompt, including previous interactions and the overall task.

Prompt patterns

Common patterns used in prompt engineering

• Instructional pattern

• Content prompt

• Style prompt

• Creative prompt

• Persona prompt

Instructional pattern

Clearly state the task you want the AI to perform

“Write a news article about…”

Content prompt

Provide specific details or references to guide the AI’s content generation

“Write a news article about (?) in the style of The New York Times”

Style prompt

Specify the desired tone or style of the output

“Write a news article about (?) in the style of The New York Times. Make sure it reads as formal and objective.”

Creative prompt

Spark the AI's creativity with starting ideas or prompts.

“Imagine a world where the cats rule the internet.”

Persona Prompt

Craft a specific persona for the AI to embody in its response.

"Write a restaurant review from the perspective of a picky eater who loves spicy food."

What is the role of persona patterns in AI responses?

• Enhances relevance and depth

• Consistency in responses

• Improves engagement

What are the steps to applying the persona pattern?

1. Understand the persona

2. Define the scenario

3. Craft the prompt

4. Review and refine

Root Prompts

Root prompts are the core instructions you provide to the AI model that define the overall task or analysis

Root Prompt: “Describe the impact of technology on society.”

Additional prompts made: “Focus on social media.”, “Consider effects on education.”

The four cornerstones of evaluation: LARF Technique

• Logical consistency

• Accuracy

• Relevance

• Factual correctness

Intelligent Assistant

software applications that use artificial intelligence to perform tasks or services based on user commands or questions. These assistants can be both physical
and virtual.

Intelligent Assistants Key Features

• Task Automations

• Voice Recognitions

• Personalization

• Integration

Intelligent Virtual Assistants Key Features

• Conversational interface

• Context awareness

• Task management

• Learning and adaptation

How do Intelligent Assistants collect user data?

1. Wake words

2. Continuous listening

3. Voice recording

4. Data transmission (cloud servers)

How virtual assistants process voice commands

1. Voice recognition

2. Natural language processing (NLP)

3. Machine Learning Algorithms

4. Task executions

Key features of Amazon Alexa

1. Voice recognition (wake word = “Alexa”)

2. Natural language processing (NLP)
   understands and interprets

3. Task automation
   sets alarms, plays music, creates shopping lists

4. Smart home integration
   example: turning off lights

5. Information retrieval
   answers questions using the web

6. Skills

7. Personalization

Amazon transcribe

(ASR) service that makes it easy for developers to add speech to text capability to their applications.

Amazon Weather Service API

The Amazon Forecast Weather Index

Amazon Polly

Uses deep learning technologies to synthesize natural-sounding human speech, so you can convert articles to speech.

With dozens of lifelike voices across a broad set of languages, use Amazon Polly to build speech-activated applications.

NLP components

Syntax: grammar rules

Semantics: the meaning of words and sentences

Speech recognition process

1. Audio input

2. Preprocessing: clean up audio

3. Feature extraction: Alexa picks out important sounds

4. Decoding: translates to text

Ontologies

An ontology is a formal representation of a set of concepts within a domain and the relationships between those concepts.

Knowledge graph

A knowledge graph is a network of real-world entities and the relationships between them, stored in a graph database.

Everyday student life tasks with virtual assistants

• Language translation

• Recipe suggestion

• Travel planning

• Event scheduling

Which tool is used to process natural language in Alexa?

Amazon Lex

AI decision-making pipeline

1. Data generated

2. Data stored

3. Data processing

4. Actionable insights

Key components of AI

Structured data

can be displayed in rows and columns (like a spreadsheet)

Unstructured data

cannot be displayed in rows and columns (like videos and images)

Data processing 

Data processing is a preparatory step in the AI model-building process. It involves cleaning, transforming, and organizing data to ensure it is suitable for analysis by AI algorithms. 

Supervised learning

Supervised learning is like a teacher guiding you to solve problems. You learn from examples that already have answers.

Unsupervised learning

Unsupervised learning is like exploring a new place without a map. You find patterns and make sense of things on your own.

What is a strategy?

A set of actions designed to achieve a long-term or overall aim.

UAE’s vision for AI.

• To become the world’s most prepared country for AI

• One of the leading nations in AI by 2031 in alignment with the UAE centennial 2071 generating AED335 billion in growth

• Boost government performance in all levels

• First in the field of AI investments

Who is UAE’s minister for AI?

Omar Sultan Al Olama

• Minister of state for AI in October 2017

• Minister of State for Artificial Intelligence, Digital Economy and Remote Work Applications in July 2020

What will the UAE chief executive officer (CEO) for AI be responsible for?

• Strategic planning

• Promoting best practices in AI

• Adopting AI across the entity’s units and projects

• Acting as an AI advisor

• Establishing government frameworks to support AI adoption

• Raising awareness

• Enhance individuals’ AI competency

UAE council for artificial intelligence and blockchain

UAE Council for Artificial Intelligence and Blockchain is tasked with proposing
policies to create an AI-friendly ecosystem; to advance research in the
sector; and to promote collaboration between the public and private sectors, including international institutions, to accelerate the adoption of AI.

National programs for AI:

1. Bachelors degree in AI

2. UAE AI internships

3. UAE AI camp

AI and advanced technology counsil

• Develop financing, investment and research plans for AI

• Foster Abu Dhabi’s strategy to transform the emirate into a hub for investments, partnerships and expertise in AI and advanced technology

• Identify investment priorities and mechanisms for research and development

• Initiate, implement and activate investment and research initiatives in AI and advanced technology

AI education and training

1. Mohamed bin Zayed’s university of AI (MBZUAI)

2. Government training programs for government employees

3. UAE open AI day

Jais Arabic LLM

UAE-US AI partnerships

• Nvidia

• AMD

• Microsoft

• OpenAI

• IBM

AI ethics

Guiding principles to ensure artificial intelligence technology is developed
and used responsibly.

Privacy concerns

1. Loss of privacy

2. Unwanted profiling

3. Targeted manipulation

4. Data misuse

Other issues

• Disinformation

• Bias

Disinformation Vs Misinformation

Misinformation is false information spread without the intent to deceive.

Disinformation is knowingly false information deliberately created and shared with the intent to mislead, manipulate, or harm.

What is academic integrity?

• Honesty: Being truthful in all academic work

• Trust: Creating confidence in the reliability of academic work.

• Fairness: Applying clear, consistent, and impartial standards.

• Respect: Valuing the learning process, ideas, and contributions of others

• Responsibility: Being accountable for one’s actions and their consequences.

• Courage: Standing up for integrity even in difficult situations.

How AI supports honesty

• Grammar and plagiarism checking

• AI-generated citations and referencing to encourage accurate attributions

• Writing aid

• Transparency reports

How can AI support fairness?

• AI-powered accessibility tools

• AI analytics for progress tracking

• use of AI and authenticity checkers

• Open AI policies between teachers and students

How does AI support respect?

• Ai can teach proper citation and respect for intellectual property

• Language tools help non-native speakers express clearly

• Ai summarization can highlight instead of steal

• Respectful AI policies acknowledge creators’ rights and data privacy.

How AI supports responsibility

• Ai can flag missing citations or incorrect sources

• Students can use AI as pure aid

• Institutional guidelines reinforce accountability

How AI supports courage

• Detecting and correcting AI misuse shows moral strength

• Discussing Ai limitations openly takes courage

• Reporting AI-assisted dishonesty promotes an ethical culture.

• Learning to question AI’s authority builds intellectual independence.

How do we evaluate AI? (ROBOT)

AI tools for research

• Aisha, ZU library

• Concensus

Emerging robot applications in manufacturing

• Collaborative robots (cobots): working alongside human works

• Robotic welding and assembly