Get detailed and easy-to-understand notes on Unit 1: Revisiting AI Project Cycle & Ethical Frameworks for AI. Learn AI project stages, ethical principles, examples, and key concepts from the Class 10 Artificial Intelligence syllabus. Perfect for CBSE board exam preparation.
Topics Covered
Revisiting AI Project Cycle & Ethical Frameworks
Welcome to Unit 1 – Revisiting AI Project Cycle & Ethical Frameworks of Class 10 Artificial Intelligence. In this unit, you will explore the AI Project Cycle, its key stages, and how ethical decision-making plays a vital role in Artificial Intelligence.
These comprehensive notes cover every topic in simple language with real-life AI examples, helping you understand how data, modeling, and evaluation come together in AI projects. You will also learn about ethical frameworks for AI, including fairness, transparency, and accountability — essential for responsible AI development.
These Class 10 AI notes are perfect for CBSE exam preparation, classroom learning, and practical AI project work. Let’s begin now!
Introduction to Project Cycle
When we are doing any work, we are following a step by step approach to do the work smoothly and in organized manner. The step by step approach helps to accompish the task and clear the steps in out mind.
Similarly, In AI model or AI project also this approach is used. It is called AI Project cycle. This project cycle follows a specific framework to achieve the goal. The AI project cycle has 6 main steps as follows:
Stage 1 Problem Scoping
This is the first stage of AI project cycle which states the problem which is going to be solved using AI model. In problem scoping various parameters are inspected which affects the problem. By performing this step the picture becomes more clear and precise.
Stage 2 Data Acquisition
Data acquisition is a process of collecting desired data for the model. The data should be collected from the authentic and valid source. There are various methods to collect data may be involved in data acquisition.
Stage 3 Data Exploration
Once data aquired, it may be in huge amount. Hence, data can be visualized to understand it in better way. Data can be visualized in forms of various graphs, flow charts, maps etc. Data exploration makes it easy to intepret the patterns in acquired data.
Stage 4 Modelling
After data exploration next stage is modelling. This stage designs a model after exploring the patterns. This model is developed to achieve the goal.
Stage 5 Evaluation
After modelling the next stage of AI Project cycle is evaluation. In evaluation, the model is tested for its efficency and accuracy. If the model is not efficient can be reviewed. The efficient model is now the base of AI project and ready for the deployment.
Stage 6 Deployment
The final stage of AI project cycle is Deployment. It is very crucial stage to check the succefull integeration of AI project. The deployed project is ready to implement in real world environments, enabling model to deliver value and impact to users and stakeholders.
Intorudction to AI Domains
The AI domains are broadly categorized into 3 domains.
Statistical Data
It is reated to data systems and processes where it collects numerious data, maintains data sets and derives meaning/sense out of them. The information extracted thorugh statistical data can be used to make a decision about it. Statistical data forms the backbone of decision-making in almost every field today.
Example of Statistical Data
- Educational institutions use statistical data to analyze students’ performance across subjects, identify learning gaps, and plan targeted interventions.
- In the business world, companies rely on sales figures, market trends, and customer feedback data to make informed decisions.
- Social media platforms like Facebook and Instagram use statistical data to track user engagement, measure ad effectiveness, and recommend personalized content.
- It is also used in predicting weather patterns
- It is also helpful in managing traffic systems.
- Statistical data also help in analyzing sports performance.
- The statistical data helps transform raw numbers into meaningful insights that guide actions and strategies.
Computer Vision
Computer Vision (CV) is a domain of Artificial Intelligence that enables machines to interpret and analyze visual information such as images and videos. It involves processes like image acquisition, screening, analysis, identification, and information extraction. Through these steps, computers learn to understand visual content and make informed decisions.
The input for CV systems can come from various sources, including cameras, sensors, and thermal or infrared devices. The goal of computer vision is to convert visual data into meaningful, computer-readable information to support automated decision-making.
Examples of Computer Vision
Agricultural Monitoring
- Computer Vision plays a vital role in modern agriculture by enabling crop monitoring, pest detection, and yield estimation.
- Using drones equipped with high-resolution cameras, farmers can capture aerial images of their fields.
- These images are then analyzed by AI systems to assess plant health, identify pest-affected areas, and estimate overall yield.
- This helps farmers make informed decisions about irrigation, fertilizer use, and harvesting schedules, ultimately improving productivity and sustainability.
Surveillance Systems
Computer Vision is widely used in surveillance to keep an eye on public areas, buildings, and borders. It helps detect unusual or suspicious activities, track people or vehicles, and send real-time alerts to security teams. With this technology, monitoring becomes faster, more accurate, and less dependent on manual observation, enhancing overall safety and security.
Few more examples of Computer Vision are: Autonomous Vehicles, Facial Recognition, Healthcare and Medical Imaging, Retail and E-commerce, Manufacturing and Quality Control, Augmented & Virtual Reality, Smart Cities, Sports and Entertainment etc.
Natural Language Processing
Natural Language Processing (NLP) is a part of Artificial Intelligence (AI) that teaches computers to understand and work with human language.
- Natural Language is the way people naturally speak and write (like English, Spanish, etc.).
- NLP uses computer rules (algorithms) to read, listen to, and figure out the meaning of these human words.
The main goal is to let computers make sense of what we say and write so they can be helpful, for example, by summarizing text, answering questions, or translating languages.
Examples of NLP
- Email filters are a common application of NLP.
- They analyze the text content (subject and body) of incoming emails to determine if the message is legitimate or spam.
- By classifying words, patterns, and structure, the filter can automatically sort and route the email, ensuring only relevant messages reach your inbox.
Machine Translation
- It automatically converts text or speech from one language to another.
- It enables people to communicate and access information across language barriers.
- Modern MT systems use deep learning models such as Neural Machine Translation (NMT), which analyze the context and meaning of sentences rather than translating word by word.
- Examples include Google Translate, Microsoft Translator, and DeepL.
- Machine translation is widely used in global communication, education, business, and online content localization, making multilingual information accessible to everyone.
Few more examples of NLP are as follows:
Ethical Framework of AI
Frameworks are simply a set of steps or rules that give you a structured plan for solving a problem.
Here’s the breakdown:
- What they are: A step-by-step guide that helps you solve problems in an organized way.
- What they do: They make sure you consider everything important and approach the problem consistently.
- Why they are useful: They help teams communicate and collaborate better by using a shared method.
You’ve probably used them without realizing it!
AI Framework Example
A common framework you’ve encountered in your AI journey is the Machine Learning Project Lifecycle (or CRISP-DM).
It’s the standard sequence of steps used to build any functional AI model:
- Define the Problem: Clearly stating what you want the AI to achieve.
- Collect and Prepare Data: Gathering, cleaning, and formatting the information.
- Choose a Model: Selecting the right algorithm (like a neural network or decision tree).
- Train the Model: Teaching the model using the prepared data.
- Evaluate the Model: Testing it to see how well it performs.
- Deploy the Model: Putting the model into a real-world application.
Ethical Framework
Ethical frameworks are like step-by-step guides that help you make sure your choices are morally sound.
- They provide a structured method for dealing with difficult moral dilemmas.
- By using them, individuals and organizations can systematically consider different ethical principles and perspectives.
- This ensures decisions are well-informed, align with core values, and avoid causing unintended harm to anyone involved.
Why do we need Ethical Frameworks for AI?
We need Ethical Frameworks for AI because AI is increasingly used for decision-making and influence, which can lead to unintended, harmful outcomes (like biased hiring algorithms).
- AI is a Decision Tool: AI is making important choices or giving advice (e.g., in hiring, lending, healthcare).
- Need for Morality: We must ensure AI’s recommendations are morally acceptable and fair.
- Preventing Harm: Ethical frameworks provide a structured way to build AI solutions that avoid bias and other negative consequences before they occur.
Factors that influence the decision making:
Types of ethical framework
The ethical framework is classified into the following types:
The thical framesworks for AI is broadly categorised into two main categories:
- Sector Based
- Values based
Sector based framework
This category focuses on specific ethical codes developed for a particular industry or domain. In other words, Sector-based ethical frameworks are industry-specific rules addressing AI’s use in particular domains. These frameworks address issues unique to that sector.
Core Idea: Ethics tailored for the unique challenges of a specific field (e.g., healthcare, finance, or law enforcement).
Example : Bioethics – for application in healthcare: This is a classic example that deals with the ethical issues arising from advances in biology and medicine. An AI system used for diagnosing cancer must adhere to the bioethical principle of autonomy (ensuring patient consent) and beneficence (acting in the patient’s best interest).
Valuse-based Framework
This is an overarching category that groups together the philosophical ethical theories focused on universal human values, rights, or consequences. This is the parent category for following:
- Rights-Based
- Utility-Based
- Virtue-Based
Rights-Based
This framework, rooted in Deontology, asserts that certain actions are inherently right or wrong, regardless of the consequences, because they uphold or violate fundamental human rights.
- Example: An AI-powered facial recognition system must respect the right to privacy. Its design should prioritize the right of individuals not to be constantly monitored, even if broader surveillance could potentially increase public safety.
Utility-Based
This framework, also known as Utilitarianism, dictates that the most ethical choice is the one that maximizes overall happiness or “good” and minimizes overall suffering or “harm” for the greatest number of people.
- Example: In a self-driving car crash scenario, a utilitarian algorithm would be designed to make a decision that results in the fewest possible injuries or fatalities, even if that means sacrificing the driver (e.g., swerving to hit a barrier instead of a bus full of people).
Virtue-Based
This framework, known as Virtue Ethics, focuses not on rules or outcomes, but on the character and moral disposition of the agent (in the context of AI, this relates to the designers and operators). It asks what a “virtuous” person would do.
- Example: An AI development team operating under a virtue framework would prioritize the virtues of honesty and transparency. They would openly disclose the limitations and potential biases of their AI model, rather than trying to obscure flaws to quickly bring a product to market.
Let’s discuss Bioethics in detail:
Bioethics
- Bioethics is the ethical framework for healthcare and life sciences.
- It ensures that AI applications in medicine and biology adhere to established ethical standards concerning health and medical issues.
Bioethics is a field of study and practice concerned with the ethical issues arising from advances in biology, medicine, and life sciences.
It is primarily guided by four foundational principles (known as the Beauchamp and Childress principles), which are used to analyze and resolve moral dilemmas in healthcare and research:
1. Autonomy
- Meaning: Respecting the patient’s right to self-determination. Individuals have the right to make informed, voluntary decisions about their own medical care, without coercion or controlling influence.
- Key Concept: Informed Consent—a direct application of this principle, ensuring the patient understands the risks, benefits, and alternatives of a procedure before agreeing to it.
2. Non-Maleficence
- Meaning: The obligation to “do no harm.” This is the oldest principle, often summarized by the Latin phrase, primum non nocere.
- Key Concept: A healthcare professional must avoid providing ineffective treatments or causing undue risk, injury, or suffering to the patient.
3. Beneficence
- Meaning: The obligation to act for the benefit of others. It requires healthcare providers to actively promote the patient’s well-being and take positive steps to prevent or remove harm.
- Key Concept: Balancing the potential benefits of a treatment against the potential risks and costs, with the goal of maximizing the good outcome for the patient.
4. Justice
- Meaning: The obligation to ensure fairness in the distribution of benefits, risks, and costs. In healthcare, it primarily relates to distributive justice—the equitable allocation of scarce resources and equal access to care.
- Key Concept: Treating patients in similar situations similarly, regardless of their background, social status, or other arbitrary factors.
The process of building AI can no longer be seen as a linear project cycle ending at deployment. Instead, the “AI Project Cycle” must be viewed as an ethical lifecycle, where human values are deliberately and constantly integrated.
We’ve established that every stage—from the initial Data Collection (addressing Bioethics and Justice) to the Modeling (addressing Utility and Non-Maleficence) and final Deployment (addressing Rights and Autonomy)—introduces new moral risks. True responsibility is achieved not by reacting to ethical failures, but by anticipating them.
The mandate for organizations is clear: Embed these ethical frameworks directly into your workflow. Treat fairness, transparency, and accountability as non-negotiable technical specifications, not as optional afterthoughts. Only by doing so can we ensure that the transformative power of AI is harnessed to serve human flourishing rather than merely corporate efficiency.
The ethical governance of AI is a continuously evolving process, not a final destination. That’s all from my side for Revisiting AI Project Cycle & Ethical Frameworks.
- 💬 COMMENT BELOW:
- At which stage of the AI Project Cycle—Data, Modeling, or Deployment—do you believe ethical risk is greatest, and what specific action is needed to mitigate it?
- 👍 LIKE this post if you believe ethical frameworks should be mandatory in all AI development.
- 🔗 SHARE this article with your classmates to drive the necessary shift from unethical to integrated AI design!
Thanks for reading this article!!!