ai project cycle ethical frameworks
NOTES
PART 1: AI PROJECT CYCLE
📌 What is AI Project Cycle?
The AI Project Cycle is a six-stage structured process used to develop an Artificial Intelligence solution.
It provides a systematic framework that helps:
- Identify a real-world problem
- Collect and analyze data
- Build and test AI models
- Deploy solutions in real environments
It ensures:
- Clear project goals
- Proper use of data
- Efficient model development
- Practical implementation
The Six Stages of AI Project Cycle
1️⃣ Problem Scoping
What is it?
This is the first and most important step.
Here, we:
- Identify the real-world problem.
- Define what we want to achieve.
- Understand the factors affecting the problem.
Key Questions to Ask:
- What is the exact problem?
- Who is affected?
- What will success look like?
- What parameters influence the problem?
Example:
If we want to predict student performance:
Parameters: attendance, study hours, previous marks.
👉 Without clearly defining the problem, AI cannot work effectively.
The Six Stages of AI Project Cycle
1️⃣ Problem Scoping
This is the first and most important stage.
- Define the problem clearly
- Identify what you want to solve using AI
- Understand factors (parameters) affecting the problem
- Set measurable goals
👉 Example: Predicting student performance based on attendance and study hours.
2️⃣ Data Acquisition
In this stage:
- Data is collected from reliable and authentic sources
- Data becomes the foundation of the project
- The quality of data determines model performance
👉 Sources may include surveys, sensors, websites, databases, etc.
3️⃣ Data Exploration
Here, we:
- Analyze collected data
- Use graphs, charts, and databases
- Identify patterns, trends, and relationships
This helps us understand:
- Missing values
- Errors in data
- Important features
👉 Example: Using bar graphs to see exam score patterns.
4️⃣ Modelling
In this stage:
- Choose a suitable AI model
- Train the model using data
- Test multiple models
- Select the most efficient one
Common AI models:
- Classification models
- Regression models
- Neural networks
5️⃣ Evaluation
Now we:
- Test the model on new (unseen) data
- Measure performance
- Identify errors
- Improve accuracy
Evaluation helps determine whether the model is ready for real-world use.
6️⃣ Deployment
This is the final stage.
- Integrate the AI model into real-world systems
- Make it available for users
- Deliver practical value
👉 Example: Deploying a chatbot on a website.
PART 2: AI DOMAINS
AI models are grouped into three major domains based on the type of data they use.
1️⃣ Statistical Data Domain
This domain:
- Works with structured data
- Uses numbers and large datasets
- Finds patterns for decision-making
👉 Examples:
- Price comparison websites
- Sales prediction systems
2️⃣ Computer Vision (CV)
Computer Vision gives machines the ability to:
- Understand images
- Analyze videos
- Extract visual information
Process includes:
- Acquiring images
- Screening
- Feature extraction
- Decision-making
👉 Examples:
- Agricultural monitoring
- Surveillance systems
- Face recognition
3️⃣ Natural Language Processing (NLP)
NLP focuses on:
- Interaction between humans and computers
- Understanding human language
- Reading, interpreting, and generating text
👉 Examples:
- Email spam filters
- Chatbots
- Google Translate
- Voice assistants
PART 3: ETHICAL FRAMEWORKS FOR AI
📌 What are Ethical Frameworks?
Ethical frameworks are:
- Structured guidelines
- Decision-making principles
- Tools to avoid unintended harm
They are especially important in AI because:
- AI systems influence decisions
- AI can impact society
- AI may affect human rights
📘 TYPES OF ETHICAL FRAMEWORKS
Ethical frameworks are divided into two major types:
1️⃣ Sector-Based Frameworks
These are designed for specific industries.
👉 Example: Bioethics
Used in:
- Healthcare
- Medical research
- Life sciences
Addresses issues like:
- Patient privacy
- Data security
- Medical fairness
2️⃣ Value-Based Frameworks
These focus on general moral principles.
They are divided into three types:
🔹 Rights-Based Framework
- Protects human rights
- Prevents discrimination
- Ensures dignity and fairness
AI must not:
- Violate privacy
- Discriminate against groups
🔹 Utility-Based Framework
- Focuses on maximum overall good
- Weighs benefits against harm
- Chooses action that benefits the majority
🔹 Virtue-Based Framework
- Focuses on character and intention
- Encourages honesty, integrity, responsibility
- Promotes ethical behavior in developers
Principles of Bioethics (Healthcare AI)
Bioethics is mainly used in healthcare and life sciences.
It is based on four core principles:
1️⃣ Respect for Autonomy
- Users must understand how AI works
- Data should be accessible and reproducible
- Transparency is essential
2️⃣ Do Not Harm (Non-maleficence)
- Avoid negative consequences
- Minimize risks
- Ensure datasets reduce harm equally
3️⃣ Ensure Maximum Benefit (Beneficence)
- Promote well-being
- Actively improve lives
- Use unbiased datasets
Not just avoiding harm, but doing good.
4️⃣ Give Justice
- Fair distribution of benefits
- Equal treatment
- No discrimination based on background
- All stakeholders must be treated fairly
Conclusion
This chapter covers:
- ✅ AI Project Cycle (6 stages)
- ✅ AI Domains (Statistical Data, CV, NLP)
- ✅ Ethical Frameworks (Sector-based & Value-based)
- ✅ Principles of Bioethics