Page 6 - Ai V2.0 Flipbook C8
P. 6
INSIDE THE SERIES
The key features of the series have been designed to ensure better
learning and assessment.
Learning Resources
02 STAGES OF AI
Learning Outcomes: PROJECT CYCLE Brainy Fact:
It provides an overview of the It presents an interesting fact relevant
unit/chapter contents. to the topic or the chapters.
Learning Outcomes
Brainy Fact
• Overview of Stages of AI Project Cycle • Problem Scoping A robot built in the late 1990s called Kismet could recognise emotions through human body
• Data Acquisition • Data Exploration language and voice tone.
• Modelling • Evaluation—The Testing Stage
• Deployment
Ethical Issues with AI
Let us now understand different ethical issues using AI.
I was reading about the stages of the AI project cycle,
and it seems quite interesting.
AI Game: That sounds exciting! What happens after that? Bias and Discrimination
At a Glance:
AI systems can inherit biases present in their training data, leading to unfair outcomes. For
It contains an interesting game or activity for example, facial recognition software has been found to have higher error rates for people with
It is! The AI project cycle involves different stages, and each
•
Wait for the AI’s response: After you play a note or a tune, the AI listens and plays its own
one plays a crucial role in building a successful AI model. This section provides a summary
After modelling, we move on to evaluation, where we test
the students to try on their own or with their darker skin tones. It is crucial to ensure AI systems are trained on diverse and representative data
part right after you.
the model's performance. Finally, the deployment stage
to avoid discrimination.
puts the model into real-world use. of the chapter.
•
Try different notes: Try playing different notes and rhythms to see how the AI adapts to
classmates to learn how the game mechanics
Can you explain the stages to me? I’m a bit confused
your music.
about how it all comes together. Privacy Violations
work while having fun. Wow, so there’s a clear step-by-step process. AI systems often require large amounts of personal data. Without proper safeguards, this data
•
Make you duet: You can keep playing, and the AI will keep responding, helping you create
At a Glance
a beautiful duet. can be misused, leading to privacy breaches. It is important to implement strong data protection
Sure! First, there's the problem scoping stage, where you measures and ensure transparency in data usage.
• AI project cycle offers a structured roadmap for creating effective AI solutions.
define the problem you want to solve. After that, you move 21 st Century #Technology Literacy • Problem scoping involves identifying a real-world issue that can be addressed using AI.
Skills
to data acquisition to gather the relevant data. Exactly! Each stage builds on the previous one to create an
• The Problem Scoping phase includes identifying the problem, understanding the context,
GAME A.I. Duet Lack of Transparency
effective AI solution.
defining goals and objectives, identifying stakeholders, setting project boundaries.
Okay, and what comes next? Many AI systems operate as “black boxes”, meaning their decision-making processes are not
About the Game
• The 4Ws Problem Canvas is a helpful tool used to break down and understand the problem step
by step.
Let’s learn more about how the Stages of AI Project Cycle easily understood. This lack of transparency can lead to mistrust and accountability issues.
A.I. Duet listens to the notes played by us and based on all the music it has heard in the Developing explainable AI is essential for building trust.
come together in practice. • The problem statement template helps us to put together all the key points into a single template
past, and responds by playing its own notes. for future reference.
Next, it's the data exploration stage. You need to clean Job Displacement
• Different problems require different types of data, and each data type brings unique value and
How to Play?
and analyse the data to understand its patterns. Then, we challenges.
proceed to modelling, where we build the AI model. Automation through AI can lead to job losses in certain sectors. While AI can create new
Step 1: Visit https://experiments.withgoogle.com/ai/ai-duet/view/ OR scan
• Data for any project, including AI projects, can be collected from two main sources: Primary and
the QR code to play the game. opportunities, it is important to consider the social impact and provide support for affected
Secondary sources.
Step 2: Click on the Play button to begin. A virtual Piano will appear on your screen. workers. • AI modelling techniques can be broadly classified into two approaches, namely, rule-based and
learning-based.
26 Touchpad Artificial Intelligence (Ver. 2.0)-VIII Security Risks
AI Innovators: AI systems can be vulnerable to attacks, such as data poisoning or adversarial inputs, which can
manipulate their behaviour. Ensuring robust security measures is vital to protect AI systems from
It presents information about the malicious activities. Video Session:
Tick ( ) the correct option.
pioneers in the field of AI. Autonomous Weapons
This section contains a link of the
1. When data is labelled, which learning technique is used?
The development of AI-powered weapons raises ethical concerns about accountability and the
Step 3: Press the Piano keys to create your own music. The notes you play will appear in a. Supervised learning b. Unsupervised learning
potential for misuse in warfare. It is important to establish international regulations to govern
blue colour. the use of such technologies. video related to the topic for better
c. Reinforced learning
d. None of these
Step 4: The AI will respond to your music with by yellow colour notes. 2. Which AI technique learns by making decisions and receiving feedback in the form of rewards or
penalties? understanding of the concept.
a. Supervised learning b. Reinforced learning 55
AI Ethics
c. Unsupervised learning d. Rule-based AI 21 st Century #Media Literacy
Skills
3. The deployment of an AI model happens after which stage?
Video Session
CEO of Google and Alphabet, Sundar Pichai has played a. Evaluation b. Modelling
a key role in making AI a core part of Google’s services. Watch the video of "A.I. Experiments: A.I. Duet" at the given link:
https://www.youtube.com/watch?v=0ZE1bfPtvZo&t=5s OR scan the QR
Under his leadership, Google AI has developed powerful c. Data exploration d. Problem scoping
code and answer the following question:
technologies like Google Assistant, Google Lens, and 4. What is the first step in the AI project cycle?
Data Collection a. Data Acquisition b. Problem Scoping
BERT (Bidirectional Encoder Representations from
This game helps you understand how AI (artificial intelligence) can work with music. You’ll What does A.I. Duet use to play a duet with the computer?
Data collection is the process of gathering information from various sources to gain insights, c. Modelling d. Evaluation
Transformers), which helps improve search results using
see how your music and the AI’s music come together, showing how AI can help make
deep learning. Pichai has also overseen Google’s expansion
make decisions, or solve problems. It can involve different methods such as surveys, interviews, 5. Which of the following data types is NOT usually collected during Data Acquisition?
music more creative.
in AI research, including self-driving technology, smart
observations, and experiments. The data collected can be either primary, gathered firsthand by a. Textual Data b. Numerical Data
home devices, and advanced machine learning tools.
c. Visual Data
the researcher, or secondary, obtained from existing sources like books, websites, or reports. Art—Stylize d. Hypothetical Data
36
Touchpad Artificial Intelligence (Ver. 2.0)-VI
Accurate and reliable data collection is essential as it forms the foundation for analysis and AI tools for art help users create and stylize images by adding different artistic effects. These
interpretation in any field of study or research. tools analyse the content and style of images, helping artists create unique pieces based on their
Touchpad Artificial Intelligence (Ver. 2.0)-VIII
46
Sundar Pichai chosen inputs. Unlike traditional style transfer methods, which require both a content image and
AI Glossary:
Data Visualisation a style image, Arbitrary Style Transfer allows any artistic style to be applied to an image. This
means users don't need a predefined style and can create creative transformations based on any
Data visualisation is the process of presenting data in a graphical or pictorial format to make style they desire. For example,
it easier to understand and interpret. Instead of looking at long lists of numbers, data can This section contains definition of
Activity: Image important AI terms.
be shown using bar graphs, pie charts, line graphs, histograms, and other visual tools. These
visuals help in identifying patterns, trends, and comparisons quickly. Data visualisation makes
This section offers a hands-on activity that 1
CEO of OpenAI, Sam Altman has been at the forefront
of generative AI research, leading the development
complex information more accessible and is widely used in fields like business, science, and
promotes experiential learning for students. AI =
of ChatGPT, GPT-4, and DALL·E. Under his leadership,
education to support decision-making.
OpenAI has created large language models (LLMs)
that can generate human-like text, assist in creative GLOSSARY
writing, and even help with coding. Altman’s work is Image
Output
2 Anonymity: The state of not revealing someone's identity, especially when collecting or
Activity pushing the boundaries of AI-powered automation, analysing data.
with applications in education, content creation,
Birthday Month and business intelligence. Bias: An unfair preference or prejudice in a system that can lead to incorrect or unjust outcomes
in AI decisions.
Purpose: To collect data, represent visually and analyse Working of AI in Stylize
Chatbots: AI-based programs that simulate human conversation and are often used in customer
The working process to create an image using Stylize is as follows:
Divide the class into four groups based on the following criteria: service or online help desks.
• Select a content image: Choose the image you want to transform, like a photo or a drawing.
a. The month with the most birthdays Computer Vision: It is a field of Artificial Intelligence that enables computers to understand
Sam Altman
• Select an artistic style: Instead of using a style image, you can choose any style you want
and interpret the visual world.
b. The month with the least birthdays (for example, the style of a famous artist, a painting, or a unique effect).
Confidentiality: Ensuring that personal or sensitive information is kept private and only
c. Birthdays falling between the 1st and 15th of the month • Apply the style: The AI tool analyses the content of your chosen image and applies the
shared with authorised people.
d. Birthdays falling between the 16th and the 30th/31st of the month selected artistic style to it.
Consent: Permission given by someone to use their data or involve them in a process, usually
• Combine the results: The AI combines the content and the style, producing a new image
after being informed about its use.
Activity Introduction where the content remains, but the style has been completely transformed.
116
Touchpad Artificial Intelligence (Ver. 2.0)-VII
Ethics: Moral principles that guide how technology, like AI, should be used responsibly and fairly.
• This activity is about data collection and analysis.
Expert System: An AI program that mimics the decision-making ability of a human expert by
AI Around Us
37
• In this activity, students will collect data from their classmates about their birthday months. using a knowledge base and set rules.
• They will then use Microsoft Excel (or Google Sheets or any similar tool) to create a Face Recognition: It is a way of identifying or confirming an individual’s identity using their faces.
graph and visually represent the data. Fuzzy Logic: A way of reasoning used in AI that deals with approximate values rather than only
true or false, helping machines handle uncertainty and partial truth.
• The purpose is to help students understand how to collect, visualise, and interpret
information using basic data tools. Machine Learning: A type of Artificial Intelligence (AI) that allows computers to learn from data
and improve their performance without being specifically programmed.
Each group will create a scrapbook to paste pictures, further categorise their data based Natural Language Processing: A field of AI that helps computers understand, interpret, and
on any one of the following options: season of birth, Kind of learner (Auditory / Visual / respond to human language, like voice commands or chat messages.
kinaesthetic), Intelligence they possess and career/ industry they aim for. Neural Network: A system of algorithms inspired by the human brain that helps computers
recognise patterns and solve problems like image recognition or speech processing.
Robotics: The design, construction, and use of robots to do tasks that are often repetitive,
dangerous, or difficult for humans.
Statistical Data: Information collected in numerical form that helps in analysing trends, making
predictions, and drawing conclusions.
44 Touchpad Artificial Intelligence (Ver. 2.0)-VII
System Map: A visual diagram that shows how different parts of a system are connected and how
they influence each other.
AI Glossary 115

