EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
HUMAN CENTERED AI
BAI613A
MODULE 1
1
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
Contents
Introduction ............................................................................................................ 3
1.1. Introduction to Artificial Intelligence ................................................................ 3
1.2. AI Growth, Expectations, and Fears ................................................................. 3
1.3. Challenges of Ignoring Humans in AI ................................................................ 4
1.4. Human-Centered AI (HCAI) ............................................................................. 5
1.5. Impact of Human-Centered AI ........................................................................ 5
1.6. What Makes HCAI Different from Traditional AI? ............................................... 6
1.7. Real World Examples of HCAI.......................................................................... 6
1.8. HCAI vs NON HCAI Summary .......................................................................... 7
2. People and Computers: A Comparison .................................................................. 8
2.1 Key difference between People and Computers ................................................. 9
3. Automation, AI, and Employment .......................................................................... 9
3.1. Jobs Likely to be Replaced by AI .................................................................... 10
3.2. Skills Engineers Should Learn ....................................................................... 10
3.3. Real Effects of Automation............................................................................ 11
3.4. Classroom Debate ....................................................................................... 11
3.5. Real-World Case Studies .............................................................................. 11
4. Conclusion ........................................................................................................ 12
2
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
Introduction
1.1. Introduction to Artificial Intelligence
What is Artificial Intelligence (AI)?
Artificial Intelligence (AI) refers to computer systems designed to perform tasks that
normally require human intelligence.
Key Characteristics of AI
• Automation of tasks
• Learns from data
• Prediction, recognition, and decision-making
Common Examples of AI
• Face recognition on smartphones
• Chatbots and voice assistants
• Recommendation systems (Netflix, Amazon)
• Fraud detection in banking
Why AI is Widely Adopted?
• Speed
• Efficiency
• Automation
• Accuracy
1.2. AI Growth, Expectations, and Fears
Growth of AI
Recent AI achievements have shown systems outperforming humans in complex tasks,
leading to:
• Global research activity
• Large investments
• Applications in medicine, manufacturing, and defence
Concerns and Fears
3
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
Along with excitement, AI raises concerns such as:
• Job loss
• Loss of human control
• Ethical misuse of technology
1.3. Challenges of Ignoring Humans in AI
The table provides a view of these challenges
Challenge Description Example Impact
Lack of Trust in Users do not understand how A loan is rejected by an AI
AI Systems decisions are made system with no
AI behaves like a black box explanation → customer Users reject AI
People hesitate to rely on such loses trust.
systems
Bias and Unfair AI learns from historical data, Hiring AI favoring one
Decisions which may contain bias gender or community due Unfair & discriminatory
Ignoring human oversight can to biased training data. outcomes
reinforce discrimination
Loss of Human Fully automated systems may Autonomous systems
Control take critical decisions making medical or legal
Dangerous decisions
No human intervention in decisions without
emergencies doctor/judge involvement
Ethical and Machines lack empathy, moral AI deciding priority of
Moral Issues reasoning, and social patients without
understanding understanding human Social harm
Ethical decisions require human context.
judgment
Safety and AI systems can fail in Self-driving car
Reliability unexpected situations misinterpreting rare road
System failures
Risks No fallback mechanism without conditions.
human-in-the-loop
Negative Blind automation leads to job Replacing customer
Impact on displacement support completely with
Social unrest
Employment No focus on reskilling or human- chatbots → poor service
AI collaboration quality
4
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
Poor User Systems are hard to use or Complex AI dashboards
Experience confusing that normal users cannot
Low adoption
Designers focus only on understand
algorithms, not users
Accountability When AI fails, who is Wrong credit scoring
Legal & governance
Problems responsible? decision with no clear
issues
Developers, users, or machines? accountability
This highlights the need for Human-Centered AI (HCAI).
1.4. Human-Centered AI (HCAI)
What is HCAI?
Human-Centered AI is a design philosophy that keeps humans in control. AI systems
support human decision-making rather than replacing humans.
Core Principles of HCAI
• Amplify Human Abilities: AI helps humans work faster and better
• Augment Human Intelligence: AI provides insights and recommendations
• Empower Users: Users retain control and understanding of AI behaviour
• Enhance Human Performance: Improves accuracy, creativity, and efficiency
1.5. Impact of Human-Centered AI
HCAI leads to:
• Better governance of AI systems
• Ethical frameworks for innovation
• Protection of human rights and dignity
• Increased social trust and acceptance of AI
By reframing AI with a human-centered vision, HCAI:
• Encourages wider adoption
• Increases positive social impact
• Aligns technology with justice and human values
5
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
1.6. What Makes HCAI Different from Traditional AI?
(1) Process – How Systems Are Built
HCAI follows human-centered design practices such as:
• User observation
• Stakeholder engagement
• Usability testing
• Iterative refinement
• Continuous evaluation of human performance
(2) Product – What Systems Become
HCAI systems act as super-tools that:
• Amplify human abilities
• Augment intelligence
• Empower users
• Enhance human performance
1.7. Real World Examples of HCAI
Real World Example HCAI Function
Scenario
Medical Decision AI-assisted radiology AI highlights possible abnormalities in X-rays or MRIs
Support Systems tools Doctor makes the final diagnosis
Explanations provided for AI suggestions
Google Maps & Route AI suggests fastest routes using real-time data
Navigation recommendations User chooses whether to follow or ignore
Systems Clear visual explanations (traffic, accidents)
Autocorrect & Grammarly, AI suggests grammar or style improvements
Writing Assistants Microsoft Editor User decides to accept or reject suggestions
Feedback improves user learning
Banking Fraud Credit card fraud AI flags suspicious transactions
Detection Systems alerts Bank staff or customer confirms the action
Accounts are not permanently blocked automatically
6
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
AI in Hiring Resume screening AI shortlists candidates
(Human-in-the- tools with human HR reviews profiles and makes final decisions
Loop Recruitment) review Bias checks included
Voice Assistants Alexa / Google AI asks for confirmation before sensitive actions
with Confirmation Assistant Users can cancel or correct commands
Privacy settings controlled by users
Smart Driving Lane assist, collision AI warns drivers of danger
Assistance warnings Driver controls steering and braking
Systems (ADAS) AI assists, not replaces driver
AI Chatbots with Customer support AI handles routine queries
Human Escalation bots Complex issues are transferred to human agents
Clear indication when human takes over
1.8. HCAI vs NON HCAI Summary
Aspect Human-Centered AI (HCAI) Non-HCAI (Traditional AI)
Human needs, values & well-
Primary focus Efficiency, speed & automation
being
Role of humans Humans remain in control Humans often removed from decision loop
Decision making Human-in-the-loop Fully automated decisions
Transparency Explainable & understandable Black-box behavior
Trust Designed to build trust Often causes mistrust
Ethics & fairness Core design requirement Usually ignored or secondary
Error handling Human oversight & correction Errors propagate automatically
Accountability Clear human responsibility Responsibility is unclear
User experience User-friendly & inclusive Technically focused, less usable
Impact on society Supports human well-being May cause harm or bias
7
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
2. People and Computers: A Comparison
Human Characteristics
• Intelligence, emotions, and consciousness
• Creativity and imagination
• Ethical and moral judgment
• Understanding of context, culture, and society
• Learn from experience
• Handle ambiguity and unexpected situations
• May get tired or make inconsistent decisions
Computer Characteristics
• Extremely fast and accurate calculations
• Process large volumes of data
• Follow predefined rules and algorithms
• Consistent and tireless
• Learn only from provided data
• Lack emotions, empathy, and moral reasoning
8
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
Key Insight
Humans excel in judgment and ethics, while computers excel in speed and data
processing. Combining both leads to safer and more effective AI systems.
2.1 Key difference between People and Computers
Aspect People (Humans) Computers Real-World Example
A human teacher handles any classroom
Type of General, flexible Narrow, task-specific
situation; AI can only answer pre-trained
intelligence intelligence intelligence
questions
Emotions & Have emotions A nurse comforts a patient; a hospital AI
No emotions
empathy and empathy cannot emotionally support
Limited to programmed Humans create new music; AI generates
Creativity Highly creative
patterns music based on existing data
Learn from life Learn from data & A driver learns from road experience; self-
Learning method
experience algorithms driving AI learns from datasets
Can make moral Cannot reason Judge considers intent in a crime; AI only
Ethical judgment
decisions ethically applies legal rules
Context Understand social Limited contextual Human interviewer adapts questions; AI
understanding & cultural context understanding follows fixed scripts
Handle
Fail in unseen Human handles sudden road block; AI may
Adaptability unexpected
situations misinterpret it
situations
Human takes minutes to calculate;
Speed Slower processing Extremely fast
computer calculates in milliseconds
Can be Human grading may vary; automated
Consistency Always consistent
inconsistent grading is uniform
Responsible for Doctor accountable for treatment; AI is not
Accountability No responsibility
decisions legally responsible
3. Automation, AI, and Employment
Key Questions
• Will AI take away jobs?
• Will robots replace humans?
• Are engineering jobs at risk?
9
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
Historical Perspective
Change is inevitable:
• Industrialization
• Computerization
• Automation
3.1. Jobs Likely to be Replaced by AI
Replaceable Tasks / Jobs
• Repetitive tasks → Data entry
• Rule-based work → Assembly line operations
• Data-intensive tasks → Basic customer queries
• Predictable tasks → Simple quality inspection
Non-Replaceable Jobs
• Creativity → Consulting, mentoring
• Emotional intelligence → Healthcare decisions
• Ethical judgment → Design and innovation
• Context understanding → Leadership and negotiation
Important Clarification
AI does not replace jobs entirely; it replaces tasks.
• Software engineers are not replaced
• Doctors are not replaced
3.2. Skills Engineers Should Learn
• AI and ML fundamentals
• Data literacy
• Automation tools
• Human-centered ethics
• Design thinking
10
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
• Communication skills
• Interdisciplinary thinking
3.3. Real Effects of Automation
Positive Effects
• Lower Costs
• Higher Quality
• Increased Demand
• Expanded Production
• New jobs and Industries
Automation creates wealth, policy and governance determines who benefits.
3.4. Classroom Debate
Topic: Will AI and automation replace jobs?
Topic: Why Unemployment is a flawed measure of Automation
Students are encouraged to discuss and debate different perspectives.
3.5. Real-World Case Studies
Case Study 1: Amazon Hiring AI Bias (2018)
What happened: Amazon developed an AI tool to screen resumes, which favored male
candidates.
Why it failed:
• Training data reflected male-dominated hiring history
• No human oversight
HCAI Lesson:
• Always include humans in the loop
• Evaluate bias and fairness before deployment
11
EAST WEST INSTITUTE OF TECHNOLOGY
(An Autonomous Institution under VTU, Belagavi)
Off Magadi Road, Bengaluru – 560091
Department of CSE(AI&ML)
Academic Year 2025-26, 6th Semester
HUMAN CENTERED AI [BAI613A] (NOTES)
Case Study 2: Apple Card Credit Limit Bias (2019)
What happened: Apple Card algorithms assigned lower credit limits to women
compared to men with similar profiles.
Why it failed:
• Historical financial data contained gender bias
• Lack of transparency in decision-making
HCAI Lesson:
• Human review is essential for sensitive decisions
• Ensure fairness, auditability, and user appeal mechanisms
4. Conclusion
Human-Centered AI ensures that technology remains aligned with human values,
ethics, and societal well-being. By keeping humans in control, HCAI creates AI systems
that are trustworthy, fair, and beneficial to society.
12