UNIT 1-Introduction

to AI
Q1: Ethical Issues in AI Development

Yes, AI development does raise ethical concerns regarding Bias, privacy, transparency, job
displacement, and autonomous weapons.

Q2: Ethical Concerns of "Home Nanny"

Privacy: 24/7 monitoring and data transmission raise privacy concerns about the child.

Security: Real-time data transmission makes the system vulnerable to hacking and data breaches.

Bias: The company's algorithms might be biased, leading to false alarms or inadequate responses.

Transparency: Parents need transparency regarding data collection and use by the company.

Dependence: Overreliance on technology could hinder parent-child bonding and discourage active
parental involvement.

Psychological impact: Constant monitoring could negatively impact the child's sense of security and
privacy.

Q3: Deployment of AI based "Smart Protector"

No, I wouldn't allow the deployment of the "Smart Protector".

Misinterpretation: The risk of misinterpreting information and launching nuclear weapons based on
false data is too high.

Loss of human control: Delegating the decision to launch nuclear weapons to AI removes human
judgment and accountability.

Escalation of conflict: The presence of such a system could encourage a more aggressive foreign
policy and increase the risk of accidental nuclear war.

Arms race: It could trigger an arms race as other countries develop their own AI-powered defenses.

Q4: Smart School Security System

Multi-layered access control: Biometrics, smart cards, geofencing for enhanced security.

Advanced monitoring & surveillance: High-definition cameras, motion sensors, smart alarms, AI-
powered analytics for real-time threat detection.

Emergency communication system: Emergency notifications, panic buttons, two-way

communication for efficient response.
Visitor management system: Pre-registration, digital logs, escort system for improved visitor
control.

Mental health & well-being features: Anonymous reporting system, mental health
resources, positive behaviour reinforcement for a safe and supportive environment.

Q5: Smartphone Data Misuse

Yes, I agree. Smartphone apps can collect and misuse data, posing threats to privacy, security, and
mental health. We need to be mindful of app permissions and data sharing.

Q6: AI Without Emotions

It might be technically possible but raises concerns. Emotionless AI could be less adaptable,
struggle with social interactions, and face ethical challenges due to its inability to understand human
emotions.

Q7: Dangerous Human Behaviours for Self-Driving Cars

Speeding: Exceeding speed limits learned from human driving data.

Aggressive driving: Tailgating, cutting off other drivers, and failing to yield the right of way.

Distracted driving: Using phones, eating, or engaging in other activities while driving.

Ignoring traffic signals: Red-light running or failing to stop at stop signs.

Reckless driving: Engaging in dangerous manoeuvres like weaving through traffic or driving under
the influence. These learned behaviours could lead to accidents and endanger other road users.

8.Smart Cities and Poverty

Smart cities can optimize resource utilization and reduce waste, leading to cost savings and
environmental benefits. Smart infrastructure can improve public transportation, safety, and
access to services, improving citizens' lives. Therefore, its justified
9.Unpredictable AI Systems
Developing AI systems that mimic human unpredictability is not desirable. While humans
exhibit complex and sometimes unpredictable behaviour, this is often due to emotional
responses, biases, and limited information processing capabilities.
AI systems should strive for:
Predictability
Transparency
Adaptability

10.Restricting AI Use Due to Hacking Risk

Restricting AI use solely due to the risk of hacking is an overly broad and potentially
counterproductive approach.
While cyber threats are a serious concern, AI can also be a powerful tool for combating
them. For example, AI can be used to:
Detect and prevent cyberattacks.
Strengthen cybersecurity defences.
Respond to cyberattacks.
11.Smart Devices for a Smarter Home
Here are a few smart devices that can make your home smarter, along with reasons for
selecting them:

1. Smart thermostat: Saves energy by automatically adjusting temperature based on your

schedule and preferences.
2. Smart lighting: Enhances convenience by allowing you to control lights remotely using
your smartphone or voice commands.
3. Smart appliances: Saves time and effort by automating tasks such as
cooking, cleaning, and laundry.
4. Smart security systems: Offers real-time alerts and notifications.
5. Smart speakers and voice assistants: Simplify tasks by controlling various smart
devices using your voice.

Q12: Smart Cities without Smart Buildings?

While smart buildings can contribute significantly to a smart city's efficiency and
sustainability, they are not strictly necessary for the concept of a smart city to exist.
Smart cities can still achieve their goals through other means, such as:
Smart infrastructure
Connected citizens.
Smart governance

Q13: Dangers of Living in Smart Cities

While smart cities offer many benefits, they also present potential dangers:
Privacy concerns: Extensive data collection and surveillance raise concerns about
individual privacy and potential misuse of personal information.
Cybersecurity risks: Smart cities are vulnerable to cyberattacks that could disrupt essential
infrastructure and services.
Technological dependence: Overreliance on technology can create inequalities,
disenfranchise vulnerable populations, and lead to a loss of control over our daily lives.
Ethical considerations: Algorithmic bias, lack of transparency in decision-making, and
potential for discrimination require careful attention and ethical guidelines.

Q14: AI in Defence Sector

Here are two examples of how AI can change the defence sector:
1. Autonomous weapon systems:
AI-powered drones and robots can perform dangerous missions with minimal human
intervention.They can improve accuracy and reduce the risk of casualties for military
personnel.
2. Enhanced situational awareness and decision-making:
AI can analyze vast amounts of data from various sources to provide real-time intelligence
and insights.This can help commanders make better decisions faster and improve battlefield
awareness.
Q15: Life without AI
Without AI, my life would be considerably different:
Daily tasks: Routine chores like setting alarms, playing music, and controlling smart devices
would require manual effort.
Information access: Research and learning would be slower and less efficient without AI-
powered search engines and information retrieval systems.
Entertainment: My access to personalized recommendations and diverse entertainment
options would be limited.
Productivity tools: Completing tasks efficiently would be more challenging without AI-
powered software for time management, communication, and collaboration.
Personalized experiences: My interactions with various services and devices would be
less tailored to my individual needs and preferences.
 UNIT 2-AI Project
Cycle
Q1: Three Parameters of Big Data
Big data is typically defined by three key parameters:
Volume: Big data refers to massive datasets containing a large amount of information. This
volume can range from terabytes to petabytes and beyond.
Velocity: Big data refers to information that is generated at a very high speed. This includes
data streams from sensors, social media interactions, and real-time transactions.
Variety: Big data refers to a wide range of data types, including structured data (tabular
format), semi-structured data (XML, JSON), and unstructured data (text, images, videos).

Q2: Unstructured Nature of Big Data

Big data is often unstructured for several reasons:
Rapid data growth: The sheer volume of data makes it difficult to store and process using
traditional methods designed for structured data.
Diverse data sources: Big data originates from various sources, including social
media, sensors, and web logs, which often lack a predefined format.
Need for flexibility: Unstructured data allows for more flexibility and adaptability to capture
a wider range of information and insights.

Q3: Best Source for AI Data Acquisition

There is no single "best" source of data for AI system data acquisition, as the ideal source
depends heavily on the specific AI project and its goals. However, some common and
valuable data sources include:
 AISHE
 UNICEF
 NCRB
 IMDB datasets
 WHO

Q4: Data Complexity and Visualization

Data complexity and data visualization are intricately related. As data complexity increases,
it becomes increasingly challenging to visualize and interpret effectively. This is because:
Non-linear relationships: Complex data often involves non-linear relationships between
variables, which are difficult to capture using traditional visualization techniques.
Hidden patterns: Identifying underlying patterns and relationships in complex data often
requires advanced visualization techniques and domain expertise.

Q5: Decision Trees and Their Creation

Here's how decision trees are created:
Data preprocessing: The data is cleaned, formatted, and prepared for analysis.
Feature selection: Relevant features are chosen based on their importance and
contribution to the model's accuracy.
Tree construction: The algorithm starts by splitting the data into smaller subsets based on a
decision rule using a chosen feature. This process continues recursively until each subset
reaches a stopping criterion, such as having all data points belong to the same class.

Q6: AI vs. Regular IT Projects

AI projects differ from traditional IT projects in several keyways:
Focus on data: AI projects are heavily data-driven, requiring significant investment in data
collection, cleaning, and processing.
Iterative development: AI models are developed and refined through an iterative process
of training, testing, and evaluation. This requires flexibility and adaptation to evolving data
and insights.
Uncertainty and ambiguity: AI models often deal with complex real-world problems with
inherent uncertainty and ambiguity.
Technical expertise: AI projects require a team with specialized expertise in data
science, machine learning, and related technical domains.

Q7: Data Requirements and Curation

Data requirements play a crucial role in data curation. The specific needs of an AI project will
dictate how data is collected.

8.Cyclicality in AI Projects
AI projects are cyclical due to:
Data-driven nature: New data triggers retraining and adjustments.
Iterative development: Experiments and feedback lead to revisiting earlier stages.
Continuous improvement: Technology advancement necessitates model updates.
Real-world feedback: Deployment reveals limitations and areas for improvement.
Ethical considerations: Ongoing evaluation and adjustments ensure responsible AI
development.

UNIT 3-ANN
Q1: Inspiration for Artificial Neural Networks (ANNs):
The idea behind ANNs stems from the observation of the human brain's structure and
function. The brain consists of billions of interconnected neurons that transmit information
through electrical signals. ANNs aim to mimic this structure by creating a network of artificial
neurons, each capable of receiving and processing information from others.

Q2: Challenges in Replicating Human Brains:

The challenges in replicating the human brain are:
Limited understanding: We still don't fully understand how the human brain
works, particularly at the cellular and network levels.
Computational limitations: Simulating the vast number of neurons and connections in the
brain requires immense computational power, exceeding current capabilities.
Individual differences: Human brains are highly individualized, with variations in structure
and function. Replicating this level of diversity is extremely difficult.

Q3: Everyday Applications of Artificial Intelligence (AI):

AI is already integrated into many aspects of our daily lives, including:
Smartphones: Virtual assistants like Siri and Google Assistant utilize AI for voice
recognition, language understanding, and personalized recommendations.
Social media: AI algorithms personalize news feeds, suggest connections, and filter content
based on user preferences.
Online shopping: AI powers product recommendations, optimizes delivery routes, and helps
prevent fraudulent transactions.
Healthcare: AI assists in medical diagnosis, analyses patient data, and personalizes treatment
plans.
Self-driving cars: AI enables autonomous driving through image recognition, obstacle
detection, and decision-making processes.
Q4: Differences between Human Brains and ANNs:
While inspired by the human brain, ANNs differ in several keyways:
Structure: Human brains are complex and intricate, with diverse neuron types and intricate
connections. ANNs, on the other hand, typically use simplified models of neurons and
connections.
Learning: Human brains learn and adapt through diverse experiences and interactions. ANNs
learn through algorithms and datasets.
Creativity and reasoning: Human brains are capable of abstract thought, creative problem-
solving, and reasoning beyond programmed algorithms. ANNs currently struggle with these
higher-level cognitive abilities.

Q5: How ANNs Improve Decision-Making:

ANNs can improve decision-making by:
Analysing large amounts of data: They can process vast amounts of data to identify patterns
and trends that humans might miss.
Making predictions: Based on data analysis.
Adapting to changing conditions: They can learn and adapt to new information, constantly
refining their decision-making abilities.