Module-1
Introduction to Artificial Intelligence (AI)
Introduction to AI
Q: What is Artificial Intelligence (AI)?
A: AI refers to the intelligence exhibited by machines or software, as opposed to human
intelligence. It involves developing systems capable of tasks like learning, reasoning, and
decision-making.
Q: What are the two main types of AI?
A:
Generative AI: Creates new data like images, text, or videos based on training data.
Example applications: Netflix, Grammarly.
1. Discriminative AI: Identifies patterns and makes predictions. Example applications:
Fraud detection, Google searches.
Q: What are some simple examples of AI?
A: Virtual assistants, recommendation systems, self-driving cars, and AI-powered medical
systems.
Q: What aspects of intelligence does AI simulate?
A: Learning, reasoning, understanding, grasping truths, seeing relationships, and separating
facts from beliefs.
Generative AI
Q: What are the tasks performed by Generative AI?
A:
1. Data Generation: Creates new content (e.g., blog posts, videos).
2. Data Transformation: Alters data for new insights.
3. Data Enrichment: Expands datasets to improve model accuracy.
Q: What are the impacts of Generative AI?
A:
Creative industries: AI-generated art and music.
Gaming: New game levels, characters, scenarios.
Natural language processing: Summaries, translations.
Discriminative AI
Q: What are the key tasks of Discriminative AI?
A:
1. Classification: Categorizing data (e.g., face unlock systems).
2. Regression: Predicting continuous values (e.g., ticket pricing).
3. Clustering: Grouping similar data points.
4. Dimensionality Reduction: Simplifying data for analysis.
5. Reinforcement Learning: Optimizing decisions through trial and error (e.g.,
gaming).
Machine Learning (ML)
Q: What is Machine Learning?
A: ML is a subset of AI that enables machines to learn and improve from data without
explicit programming.
Q: What are the types of ML?
A:
1. Supervised Learning: Trains on labelled data for specific predictions.
2. Unsupervised Learning: Identifies patterns in unlabelled data.
3. Deep Learning: Involves complex layers for tasks like image recognition.
Q: What are the steps in ML training?
A:
1. Training: Using data to build the model.
2. Validation: Ensuring the model performs well.
3. Evaluation: Testing the model with new data.
Q: What are common challenges in ML?
A: Data bias, incorrect predictions due to poor training, and high memory and processing
requirements.
AI and Human Intelligence
Q: How does AI differ from human intelligence?
A: While AI simulates human intelligence through algorithms, human intelligence involves
emotions, intuition, and consciousness developed through evolution.
Q: How can AI be defined in terms of behaviour?
A:
1. Acting humanly (e.g., CAPTCHA tests).
2. Thinking humanly (e.g., cognitive modelling).
3. Thinking rationally (logical problem solving).
4. Acting rationally (decision-making in environments).
Miscellaneous Key Points
Q: What is "AI Winter"?
A: A period of reduced funding and interest in AI research due to unmet expectations.
Q: What are the roles of algorithms in ML?
A: Algorithms are sets of rules for problem-solving or task performance. They are central to
ML for training models and ensuring accuracy.
Q: What is the role of statistics in ML?
A: It underpins several ML approaches, such as probabilistic inference and pattern
recognition.
ISO 42001 Module 1 - Key Questions & Answers
Introduction to Artificial Intelligence (AI)
Q: What is Artificial Intelligence (AI)?
A: AI is the intelligence displayed by machines, designed to simulate human cognitive
functions such as learning, reasoning, and decision-making.
Q: Who coined the term AI?
A: John McCarthy in 1955.
Q: What are examples of AI in daily life?
A: Virtual assistants (e.g., Alexa, Siri), self-driving cars, AI-powered medical tools, drones,
and recommendation systems like Spotify or Netflix.
Q: How does AI disrupt or enhance life?
A: AI introduces conveniences (like personalized assistants) but can also lead to ethical
dilemmas and societal shifts in creativity, employment, and decision-making.
Q: What are the aspects of intelligence AI tries to simulate?
A:
1. Learning: Absorbing and processing information.
2. Reasoning: Manipulating information for outcomes.
3. Understanding: Interpreting relationships and truths.
4. Fact vs. Belief: Distinguishing validated data from assumptions.
Types of AI
Q: What are the two main classes of AI?
A:
1. Generative AI: Creates new content based on learned data (e.g., artworks, music).
2. Discriminative AI: Classifies or predicts data patterns (e.g., fraud detection,
customer segmentation).
Q: What are examples of Generative and Discriminative AI applications?
A:
Generative: Netflix’s recommendation engine, AI art generation.
Discriminative: Google search algorithms, Amazon product categorization.
Machine Learning (ML)
Q: What is Machine Learning (ML)?
A: A subset of AI where computers learn and improve from data to make predictions or
perform tasks without explicit programming.
Q: What are the types of ML?
A:
1. Supervised Learning: Uses labelled datasets (e.g., beauty ratings).
2. Unsupervised Learning: Finds patterns in unlabelled datasets.
3. Deep Learning: Advanced learning with complex layers, used in image recognition
and autonomous vehicles.
Q: What are the steps in training ML models?
A:
1. Training: Building the model using datasets.
2. Validation: Ensuring the model’s accuracy.
3. Evaluation: Testing the model on new data.
Q: What is backpropagation in ML?
A: A technique where errors in predictions are used to adjust and refine the model for better
outcomes.
Generative AI
Q: What tasks can Generative AI perform?
A:
1. Data Generation: Producing new content (e.g., blogs, videos).
2. Data Transformation: Modifying existing data for insights.
3. Data Enrichment: Enhancing datasets to improve robustness.
Q: What industries are impacted by Generative AI?
A:
Creative industries: Digital artworks, music, and films.
Gaming: Content creation like levels or characters.
Natural Language Processing (NLP): Summaries, translations, and language models.
Discriminative AI
Q: What are the tasks of Discriminative AI?
A:
1. Classification: Categorizing data (e.g., face unlock systems).
2. Regression: Predicting continuous outcomes (e.g., ticket prices).
3. Clustering: Grouping data points (e.g., Spotify recommendations).
4. Dimensionality Reduction: Simplifying datasets.
5. Reinforcement Learning: Learning through trial and error to optimize results (e.g.,
gaming AI).
AI and Human Intelligence
Q: How does AI compare to human intelligence?
A: AI replicates cognitive functions like decision-making using algorithms but lacks
emotional depth, intuition, and consciousness.
Q: What are the four ways to define AI behaviour?
A:
1. Acting Humanly: Engaging in human-like interactions (e.g., Turing Test).
2. Thinking Humanly: Modelling human cognition.
3. Thinking Rationally: Logical problem-solving.
4. Acting Rationally: Decision-making based on data and environmental factors.
Important Concepts in AI and ML
Q: What is the "AI Winter"?
A: A period of stagnation in AI development due to reduced funding and failed expectations.
Q: What role do algorithms play in AI?
A: Algorithms are the backbone of AI and ML, providing rules and methods for tasks such as
pattern recognition, classification, and data processing.
Q: What are the common challenges in ML?
A:
Data biases.
Complexity in selecting algorithms.
Preparing high-quality datasets.
Balancing simplicity with model accuracy.