Introduction to Artificial Intelligence and Machine Learning
Unit – 1
FUNDAMENTALS OF ARTIFICIAL
INTELLIGENCE (AI)
26 August 2025 1
How AI can bolster smart cities: Closing
tech gaps in infrastructure
[Link]
cities-closing-tech-gaps-in-infrastructure
26 August 2025 2
AI-Driven Urban Management
• Traffic optimization: Cities like Delhi and Mumbai are deploying AI-
based traffic management systems—including ANPR cameras and
adaptive signal control—to reduce congestion and improve road.
• Crowd prediction & public safety: AI is used for crowd monitoring
during large events, predictive policing, and crime detection. In Surat,
AI-based surveillance systems helped reduce crime by 27%.
26 August 2025 3
Intelligent Waste & Maintenance Systems
• Smart waste management: AI-enabled sensors and routing help
optimize waste collection, prevent overflow, and lower operational
costs.
• Infrastructure health monitoring: Machine learning models predict
when roads, pipelines, or utilities require maintenance, reducing
breakdowns and repair costs
26 August 2025 4
Energy & Grid Intelligence
Smart grid optimization: AI enhances forecasting, load balancing, and
renewable integration, reducing blackouts and improving asset
utilization.
Edge computing & decentralization: Edge processing reduces latency
in IoT-led infrastructure—vital for responsive services and data-driven
control.
Predictive maintenance: Sensor-driven analytics detect faults early and
schedule repairs proactively across transformers, substations, and EV
charging infrastructure
26 August 2025 5
Sustainability & Urban Resilience
Heat vulnerability mapping: AI and satellite imagery help develop
building-level heat risk profiles in Delhi, allowing targeted interventions
in high-risk neighborhoods
Energy-efficient design: AI optimizes building usage of power and
resources, aligning with sustainable urban planning goals
26 August 2025 6
Skills, Governance & Ethical Imperatives
Workforce skilling: There’s a pressing need to upskill public-sector
staff in AI, data analytics, and smart-city systems to ensure effective
implementation
Policy and regulation: India is building frameworks for AI
governance—including citizen data privacy, responsible deployment,
transparency, and cybersecurity oversight
26 August 2025 7
26 August 2025 8
26 August 2025 9
Practice Question
The term "artificial intelligence" was first coined during a
conference held at Dartmouth in 1965.
• True
• False
26 August 2025 10
26 August 2025 11
26 August 2025 12
26 August 2025 13
26 August 2025 14
26 August 2025 15
Practice Question
What are the essential components of Artificial Intelligence?
• Perceive, Predict, and Control
• Sense, Comprehend, Act
• Sense, Compute, and Execute
• Learn, Sense, and Think
26 August 2025 16
26 August 2025 17
26 August 2025 18
26 August 2025 19
26 August 2025 20
26 August 2025 21
26 August 2025 22
26 August 2025 23
Practice Question
Symbolic AI was an early approach in AI development,
focusing on using symbols and rules to represent knowledge.
• False
• True
26 August 2025 24
26 August 2025 25
26 August 2025 26
26 August 2025 27
26 August 2025 28
26 August 2025 29
26 August 2025 30
Practice Question
What is one of the applications of computer vision that helps
improve operational efficiency?
• Object tracking to monitor vehicle movements
• Natural language processing for sentiment analysis
• Autonomous robots in hazardous environments
• Data classification for identifying malicious network actors
26 August 2025 31
26 August 2025 32
26 August 2025 33
26 August 2025 34
SNA - Social Network Analytics
26 August 2025 35
26 August 2025 36
26 August 2025 37
26 August 2025 38
Practice Question
The use of clusters in network analysis can provide insights
into areas such as fraud detection, supply chain management,
and cybersecurity.
• False
• True
26 August 2025 39
26 August 2025 40
26 August 2025 41
Practice Quiz
26 August 2025 42
Quiz - 1
26 August 2025 43
Quiz - 2
26 August 2025 44
26 August 2025 45
26 August 2025 46
26 August 2025 47
26 August 2025 48
26 August 2025 49
Narrow AI (Weak AI)
• Narrow AI is designed and trained on a specific task or a narrow
range tasks. These Narrow AI systems are designed and trained for a
purpose. These Narrow systems performs their designated tasks but
mainly lack in the ability to generalize tasks.
• Examples:
• Voice assistants like Siri or Alexa that understand specific commands.
• Facial recognition software used in security systems.
• Recommendation engines used by platforms like Netflix or Amazon.
26 August 2025 50
How Does Narrow AI Work?
• Narrow AI systems are trained using large datasets and sophisticated
machine learning models, allowing them to perform specific tasks
without requiring human intervention. These systems are highly
efficient at solving targeted problems but cannot apply their
intelligence beyond the designated tasks.
• For example, a Narrow AI-powered chatbot might excel at answering
customer service queries but will be unable to perform any unrelated
tasks like driving a car or diagnosing medical conditions. This lack of
versatility distinguishes Narrow AI from the broader vision of General
AI, which remains a goal for the future.
26 August 2025 51
Characteristics of Narrow AI
• Specialization: Narrow AI is built to handle a specific problem or perform a
defined task. It cannot think beyond its programming.
• Data Dependency: The effectiveness of Narrow AI relies on the quality and
quantity of the data it is trained on. The more relevant data it receives during
training, the more accurate its predictions and decisions will be.
• Lack of Self-Awareness: Narrow AI systems do not possess consciousness or
self-awareness. They cannot think or reason independently but instead follow
patterns and rules dictated by their programming.
• Reactive Nature: Most Narrow AI systems react to specific inputs based on their
training. They do not have the ability to generate new knowledge or conceptual
understanding beyond their programmed tasks.
• Task-Specific Optimization: These AI systems are optimized for efficiency and
accuracy in a particular domain. This makes them excellent at performing their
designated tasks but limits their versatility.
26 August 2025 52
Applications of Narrow AI
• Virtual Assistants: Systems like Siri, Google Assistant, and Alexa are prime examples of
Narrow AI in action. These systems can understand and respond to voice commands, set
reminders, play music, and perform other tasks — all within their programmed capacity.
• Recommendation Engines: Platforms like Netflix, Spotify, and Amazon use Narrow AI
to analyze user behavior and recommend personalized content, improving user experience
and engagement.
• Facial Recognition: Narrow AI systems are widely used in security applications like
facial recognition, where they can identify individuals based on their facial features.
• Customer Service Bots: AI chatbots are increasingly used in customer service to respond
to inquiries, provide product information, and resolve common issues. While effective,
they are limited to the scripts and data they have been trained on.
• Autonomous Vehicles: Self-driving cars, like those developed by Tesla, rely heavily on
Narrow AI to navigate roads, recognize obstacles, and follow traffic rules. However, these
systems are specialized to driving and cannot perform tasks outside this domain.
26 August 2025 53
Benefits of Narrow AI
• Efficiency: Narrow AI systems can automate routine tasks, saving
time and reducing human effort in areas such as customer service, data
entry, and supply chain management.
• Accuracy: In fields like healthcare and finance, Narrow AI provides
accurate data analysis and prediction capabilities, improving decision-
making processes.
• Cost Reduction: By automating repetitive tasks, Narrow AI helps
businesses cut costs and improve operational efficiency.
26 August 2025 54
Limitations of Narrow AI
• Task-Specific Nature: Narrow AI can only perform the tasks it was
specifically designed and trained for. It cannot adapt to new, unrelated
tasks without significant retraining or reprogramming.
• Lack of Understanding: Narrow AI lacks human-like understanding
and reasoning capabilities. It can perform well within its domain, but it
does not possess general knowledge or the ability to think beyond
predefined constraints.
• Ethical Concerns: Narrow AI systems, particularly in areas like facial
recognition, have raised concerns over privacy, bias, and surveillance,
prompting ongoing debates about ethical guidelines and regulations.
26 August 2025 55
General AI (Strong AI)
• General AI refers to AI systems that have human intelligence and
abilities to perform various tasks. Systems have capability to
understand, learn and apply across a wide range of tasks that are
similar to how a human can adapt to various tasks.
• While General AI remains a theoretical concept, researchers aim to
develop AI systems that can perform any intellectual task a human
can. It requires the machine to have consciousness, self-awareness,
and the ability to make independent decisions, which is not yet
achievable.
26 August 2025 56
Examples of General AI
• Chatbots: Chatbots use natural language processing (NLP) to analyze
human speech and generate responses.
• Customer service: General AI based customer service would access
large amount od customer data and provide personalized service in
real-time.
• Imagination: General intelligence can read and understand human-
made code and make it better.
26 August 2025 57
Potential Applications of General AI:
• Robots that can learn new skills and adapt to unforeseen
challenges in real-time.
• AI systems that could autonomously diagnose and solve
complex medical issues across various specializations.
26 August 2025 58
Superintelligence (Super AI)
• Super AI surpasses intelligence of human in solving-problem,
creativity, and overall abilities. Super AI develops emotions, desires,
need and beliefs of their own. They are able to make decisions of their
own and solve problem of its own. Such AI would not only be able to
complete tasks better than humans but also understand and interpret
emotions and respond in a human-like manner.
• While Super AI remains speculative, it could revolutionize industries,
scientific research, and problem-solving, possibly leading to
unprecedented advancements. However, it also raises ethical concerns
regarding control and regulation.
26 August 2025 59
Importance of Artificial Super Intelligence
(ASI)
• Artificial Super Intelligence (ASI) is really good at solving hard problems, like
stopping climate change or finding treatments for illnesses.
• ASI could help us create new things much faster in areas like medicine, energy,
and exploring space.
• ASI could help scientists find things quicker, like learning more about space,
genes, or new materials.
• ASI might help countries cooperate better by giving good advice and helping
make decisions that benefit everyone.
26 August 2025 60
Importance of Artificial Super Intelligence
(ASI)
• ASI could help people communicate better, even if they speak different languages
or come from different cultures.
• ASI could help us use things like energy and water more wisely, which would be
good for the environment and save money.
• ASI might help doctors find better ways to treat patients and detect diseases early,
so people can live longer and healthier lives.
• ASI makes us think about what's right and wrong when it comes to using
technology, so we need rules to ensure it's used in a good way.
26 August 2025 61
Types of Artificial Intelligence Based on
Functionalities
• AI can also be classified into four types based on how the systems
function.
• This classification is more commonly used to distinguish AI systems
in practical applications.
26 August 2025 62
Reactive Machines
• Reactive machines are the most basic form of AI. They operate purely
based on the present data and do not store any previous experiences or
learn from past actions. These systems respond to specific inputs with
fixed outputs and are unable to adapt.
• Examples:
• IBM’s Deep Blue, which defeated the world chess champion Garry Kasparov
in 1997. It could identify the pieces on the board and make predictions but
could not store any memories or learn from past games.
• Google’s AlphaGo, which played the board game Go using a similar approach
of pattern recognition without learning from previous games.
26 August 2025 63
Limited Memory in AI
• Limited Memory AI can learn from past data to improve future
responses. Most modern AI applications fall under this category. These
systems use historical data to make decisions and predictions but do
not have long-term memory. Machine learning models, particularly in
autonomous systems and robotics, often rely on limited memory to
perform better.
• Examples:
• Self-driving cars: They observe the road, traffic signs, and movement of
nearby cars, and make decisions based on past experiences and current
conditions.
• Chatbots that can remember recent conversations to improve the flow and
relevance of replies.
26 August 2025 64
Theory of Mind
• Theory of Mind AI aims to understand human emotions, beliefs,
intentions, and desires. While this type of AI remains in development,
it would allow machines to engage in more sophisticated interactions
by perceiving emotions and adjusting behavior accordingly.
• Potential Applications:
• Human-robot interaction where AI could detect emotions and adjust its
responses to empathize with humans.
• Collaborative robots that work alongside humans in fields like healthcare,
adapting their tasks based on the needs of the patients.
26 August 2025 65
Self-Awareness AI
• Self-Aware AI is an advanced stage of AI that possesses self-consciousness
and awareness. This type of AI would have the ability to not only
understand and react to emotions but also have its own consciousness,
similar to human awareness.
• While we are far from achieving self-aware AI, it remains the ultimate goal
for AI development. It opens philosophical debates about consciousness,
identity, and the rights of AI systems if they ever reach this level.
• Potential Applications:
• Fully autonomous systems that can make moral and ethical decisions.
• AI systems that can independently pursue goals based on their understanding of the
world around them.
26 August 2025 66
AI Ethics :
Challenges, Importance and Future
• Artificial Intelligence (AI) is reshaping industries, decision-making,
and daily life but with this power comes responsibility.
• As AI systems become more capable and autonomous questions about
fairness, accountability, privacy and bias grow.
• AI ethics focuses on guiding the development and use of AI in ways
that align with human values and societal norms.
26 August 2025 67
26 August 2025 68
26 August 2025 69
Bias and Fairness
• Algorithmic Bias: Algorithmic bias refers to the systematic and unfair
discrimination that can arise from the data and algorithms used in AI
systems. This bias can manifest in various ways, leading to unequal
treatment of individuals based on race, gender, age, or other attributes.
• Discrimination in AI Applications: AI applications can perpetuate and
even exacerbate existing societal biases, resulting in discriminatory
outcomes. Examples include biased hiring algorithms, racially biased facial
recognition systems, and discriminatory credit scoring models.
• Strategies for Mitigating Bias: Mitigating bias in AI requires a
multifaceted approach. This includes diversifying training data,
implementing fairness-aware algorithms, conducting regular audits, and
fostering an inclusive development environment.
26 August 2025 70
Privacy and Data Protection
• Data Collection and Consent: AI systems often rely on vast amounts of
data, raising concerns about privacy and informed consent. It is essential to
establish clear guidelines for data collection, ensuring that individuals are
aware of and agree to how their data is used.
• Surveillance and Monitoring: The use of AI for surveillance and
monitoring poses significant privacy risks. Without proper regulation, these
technologies can lead to invasive tracking and profiling of individuals.
• Ensuring Data Security: Ensuring the security of data used in AI systems
is critical to prevent breaches and misuse. Robust encryption, access
controls, and regular security assessments are necessary to protect sensitive
information.
26 August 2025 71
Transparency and Explainability
• Black Box Problem: Many AI systems operate as "black boxes," meaning
their internal workings are not transparent or understandable. This lack of
transparency can lead to mistrust and difficulties in accountability.
• Importance of Explainable AI: Explainable AI aims to make AI systems
more transparent and understandable to users and stakeholders. This is
crucial for building trust and ensuring that AI decisions can be scrutinized
and challenged.
• Approaches to Enhancing Transparency: Enhancing transparency
involves developing AI models that provide clear explanations for their
decisions, adopting standardized documentation practices, and promoting
open-source AI development.
26 August 2025 72
Accountability and Responsibility
• Who is Responsible for AI Decisions?: Determining responsibility for AI
decisions can be complex, involving developers, deployers, and users. Clear
guidelines are needed to assign accountability at various stages of the AI
lifecycle.
• Legal and Ethical Accountability: Establishing legal and ethical
accountability frameworks is essential for addressing harms caused by AI
systems. This includes regulations, ethical guidelines, and mechanisms for
redress.
• Frameworks for Responsible AI Development: Frameworks for
responsible AI development encompass principles such as fairness,
accountability, transparency, and ethics (FATE). These frameworks guide
the ethical design, deployment, and governance of AI technologies.
26 August 2025 73
Autonomy and Control
• Human Oversight in AI Systems: Human oversight is crucial to ensure
that AI systems operate within ethical boundaries and do not cause
unintended harm. This includes human-in-the-loop approaches and robust
monitoring mechanisms.
• Autonomous Weapons and Ethical Concerns: The development of
autonomous weapons raises significant ethical concerns. Ensuring human
control over lethal autonomous systems is vital to prevent indiscriminate
harm and uphold humanitarian principles.
• Balancing AI Autonomy with Human Control: Balancing AI autonomy
with human control involves setting limits on the decision-making
capabilities of AI systems and ensuring human intervention in critical
situations.
26 August 2025 74
Ethical Use of AI in Medicine and
Healthcare
• AI in Diagnosis and Treatment: AI has the potential to revolutionize
healthcare by improving diagnosis and treatment. However, ethical
considerations such as accuracy, fairness, and patient consent must be
addressed.
• Patient Privacy and Consent: Ensuring patient privacy and informed
consent is critical in AI-driven healthcare. Transparent data practices and
stringent privacy protections are necessary to maintain trust and ethical
standards.
• Ethical Considerations in AI-Driven Healthcare: Ethical considerations
in AI-driven healthcare include addressing biases in medical algorithms,
ensuring equitable access to AI benefits, and maintaining human oversight
in clinical decision-making.
26 August 2025 75
Autonomy and Control
• Human Oversight in AI Systems: Human oversight is crucial to ensure
that AI systems operate within ethical boundaries and do not cause
unintended harm. This includes human-in-the-loop approaches and robust
monitoring mechanisms.
• Autonomous Weapons and Ethical Concerns: The development of
autonomous weapons raises significant ethical concerns. Ensuring human
control over lethal autonomous systems is vital to prevent indiscriminate
harm and uphold humanitarian principles.
• Balancing AI Autonomy with Human Control: Balancing AI autonomy
with human control involves setting limits on the decision-making
capabilities of AI systems and ensuring human intervention in critical
situations.
26 August 2025 76
Job Displacement and Economic Impact
• AI and Automation in the Workforce: AI and automation are
transforming the workforce, leading to job displacement in certain sectors
while creating new opportunities in others. Addressing the economic impact
requires proactive policies and workforce retraining programs.
• Economic Inequality and Job Loss: The uneven distribution of AI benefits
can exacerbate economic inequality. Strategies to mitigate this include
social safety nets, equitable access to AI technologies, and inclusive
economic policies.
• Policies for Managing Economic Disruption: Governments and
organizations must develop policies to manage economic disruption caused
by AI. This includes education and training initiatives, support for displaced
workers, and fostering innovation in emerging industries.
26 August 2025 77
AI and Human Rights
• AI and Freedom of Expression: AI can impact freedom of expression
through content moderation and censorship. Ensuring that AI systems
respect freedom of speech while combating harmful content is a delicate
balance.
• AI in Law Enforcement and Justice: AI applications in law enforcement
and justice, such as predictive policing and risk assessment tools, raise
significant human rights concerns. Ensuring fairness, accountability, and
transparency in these systems is crucial.
• Protecting Human Rights in the Age of AI: Protecting human rights in the
age of AI involves establishing legal and ethical safeguards, promoting
international cooperation, and ensuring that AI technologies are aligned
with human rights principles.
26 August 2025 78
Ethical Challenges of AI
Some of the common AI Ethics challenges that are discussed below:
• Opacity: Many AI systems especially deep learning models function
as "black boxes." It becomes difficult to trace how a particular
outcome was generated. This lack of transparency hampers trust and
limits accountability.
• Attacks and breaches: As they rely heavily on data pipelines and
model integrity, AI models are vulnerable to data breaches and
adversarial attacks. Ensuring data confidentiality, authenticity and
robustness is crucial.
26 August 2025 79
Ethical Challenges of AI
• Algorithmic biases: Biases present in training data or algorithmic
decision-making processes can result in unfair outcomes. Such biased
data leads to an unethical AI.
• Ethical Accountability: Developers and organizations must take
responsibility for their models impact. A small error in model logic or
deployment can trigger widespread issues like privacy violations or
safety risks.
• Risk Management: AI systems can fail in unpredictable ways.
Addressing these risks requires safety measures, validation and
anticipation of consequences before deployment.
26 August 2025 80
AI Code of Ethics
• Openness and Transparency: Developers should build systems
whose decisions and processes can be understood. Logging system
decisions, model architecture and versioning are essential parts of this
effort.
• Data Security Standards: Systems should protect user data through
encryption, access control and authentication mechanisms to prevent
unauthorized leakage.
26 August 2025 81
AI Code of Ethics
• Fairness and Equity: AI systems must be evaluated to avoid biased
decisions. This includes using diverse datasets, fairness-aware
algorithms and performance testing across different demographic
groups.
• Responsibility and Accountability: Ethical responsibility lies with
individuals and organizations involved in AI development. Clear
guidelines and documentation ensure that all contributors understand
the system’s intended and unintended effects.
• Safety and Well-being: Systems should be rigorously tested for
safety. Failsafe mechanisms and rollback procedures are critical
components of safe deployment.
26 August 2025 82
Implementing Ethics of AI
Here’s how developers can implement key ethical principles during AI
development:
• Promoting Transparency: Integrate interpretable models to trace
decision logic. Maintain structured documentation of development
steps, inputs and system behavior logs.
• Ensure Security: Apply data encryption during storage and transfer.
Incorporate role-based access and authentication to restrict
unauthorized system usage.
26 August 2025 83
Implementing Ethics of AI
• Reduce Bias and Ensure Fairness: Audit datasets for class
imbalance and underrepresentation. Evaluate models using fairness
metrics like demographic parity and equal opportunity.
• Ensure Responsibility: Define clear accountability through
governance policies. Include ethical reviews and compliance checks at
key development stages.
• Strengthen Safety: Perform edge-case testing and simulate failure
scenarios. Deploy fallback mechanisms or human oversight to catch
high-risk errors.
26 August 2025 84
Steps to Make AI More Ethical
Here are some key steps to promote ethical AI development and deployment:
• Define Clear Ethical Guidelines: Establish company-wide standards for
responsible AI use.
• Debias Data and Algorithms: Use preprocessing techniques and
rebalancing to ensure fairness.
• Respect User Consent: Obtain informed consent and provide data control
options to users.
• Design for Ethics: Add ethical thinking into the entire AI lifecycle, from
ideation to post-deployment monitoring.
• Build Accountability Structures: Track and document decisions and
impacts throughout development.
26 August 2025 85
Future of AI Ethics
The landscape of AI ethics is rapidly evolving. We can expect:
• Advanced Research in Ethical AI: New frameworks for
interpretable, fair and safe AI will continue to emerge.
• Stronger Regulatory Oversight: Governments and international
organizations are actively developing laws and standards for AI
governance.
• Ethics in Emerging Domains: As AI enters new fields like
autonomous vehicles and predictive healthcare, ethical concerns will
become even more important.
26 August 2025 86
26 August 2025 87
26 August 2025 88
26 August 2025 89
26 August 2025 90
26 August 2025 91
26 August 2025 92
26 August 2025 93
26 August 2025 94
26 August 2025 95
26 August 2025 96
26 August 2025 97
26 August 2025 98
26 August 2025 99
26 August 2025 100
26 August 2025 101
