Demystifying AI: What It Is and How It Works
After this lesson you can define AI and distinguish it from machine learning and deep learning, name the types of AI and the core components of an ML system, and recognise everyday real-world AI applications.
{"contentFormat":"slides.v1","completion":{"requireAllSlides":true,"requireQuiz":true},"slides":[{"kind":"title","eyebrow":"Module 1: AI Foundations and Mindset","title":"Demystifying AI: What It Is and How It Works","body":"Before we build AI, we must understand what AI really is. This lesson clarifies core concepts, distinguishes AI from machine learning and deep learning, and grounds you in the current 2026 landscape where narrow AI and generative models dominate.","outcomes":["Define artificial intelligence and its key capabilities","Differentiate AI, machine learning, and deep learning","Identify current real-world AI applications (2026)","Explain the role of data, algorithms, and models","Recognize that only narrow AI exists today"],"narration":"Welcome to the first lesson. AI is often misunderstood. Let's demystify it together, focusing on what's real and practical in 2026."},{"kind":"content","heading":"What is Artificial Intelligence?","body":"At its core, AI is the field of creating machines that can perform tasks requiring human-like intelligence. In 2026, this is achieved through narrow AI — systems designed for specific tasks. Key capabilities include:\n\n- Learning: Acquiring information and rules from data.\n- Reasoning: Using rules to reach conclusions.\n- Problem Solving: Searching for solutions.\n- Perception: Interpreting sensor data (cameras, microphones).\n- Language Understanding: Processing human language.\n\n> Note: The dominant paradigm today is generative AI (e.g., ChatGPT, DALL-E 3) and large language models (LLMs) built on the transformer architecture with attention mechanisms. These are all narrow AI.","callout":{"variant":"insight","title":"Current State (2026)","text":"All AI in use today is narrow AI. General AI (human-level) and Super AI remain theoretical. Even advanced systems like autonomous vehicles operate at Level 2 or 3 autonomy — they are not fully self-driving."},"narration":"AI is about mimicking human intelligence. But today, all AI is narrow — it excels at one task. Generative AI and LLMs are the leading examples."},{"kind":"content","heading":"AI vs. Machine Learning vs. Deep Learning","body":"These terms are often used interchangeably, but they form a hierarchy:\n\n- AI: The broadest concept — any technique that enables machines to mimic human intelligence.\n- Machine Learning (ML): A subset of AI where algorithms learn patterns from data without explicit programming.\n- Deep Learning (DL): A subset of ML using multi-layered neural networks to model complex patterns. Foundational to modern AI.\n\nAnalogy: Baking. AI is the goal of baking something. ML is using a recipe and learning from past bakes. DL is a complex recipe with many layers, enabling intricate creations.","table":{"headers":["Concept","Scope","Example (2026)"],"rows":[["Artificial Intelligence","Broadest","Any system that mimics human intelligence"],["Machine Learning","Subset of AI","Spam filter, recommendation system"],["Deep Learning","Subset of ML","ChatGPT, DALL-E 3, self-driving perception"]]},"narration":"Think of AI as the big umbrella. ML is a part of it, and deep learning is a part of ML. Deep learning powers today's most impressive AI tools."},{"kind":"content","heading":"Key Components of Machine Learning","body":"Every ML system relies on five core components:\n\n1. Data: The foundation. Quality and quantity matter.\n2. Algorithms: Statistical methods (e.g., linear regression, neural networks).\n3. Model: The learned representation after training.\n4. Training: Feeding data to the algorithm to learn patterns.\n5. Evaluation: Testing the model on unseen data.\n\nExample: A spam filter is trained on thousands of emails (data) using a classification algorithm. The resulting model can then classify new emails as spam or not.","callout":{"variant":"tip","title":"Why This Matters","text":"Understanding these components helps you debug and improve AI systems. In 2026, most AI work involves curating data, choosing algorithms, and evaluating models."},"narration":"Data, algorithms, model, training, evaluation — these are the building blocks. Get comfortable with them."},{"kind":"content","heading":"Real-World AI Applications (2026)","body":"Here are concrete examples of narrow AI in use today:\n\n- Spam Filters: ML classifies emails as spam or not.\n- Recommendation Systems: Netflix, Amazon suggest content based on your history.\n- Generative AI: ChatGPT for text, DALL-E 3 for images, GitHub Copilot for code.\n- Facial Recognition: Smartphone unlock, airport security.\n- Virtual Assistants: Siri, Alexa, Google Assistant use NLP.\n- Medical Diagnosis: AI analyzes X-rays, MRIs to detect diseases.\n- Fraud Detection: Banks flag unusual transactions.\n- Autonomous Vehicles: Limited Level 2-3 autonomy; not fully self-driving.","callout":{"variant":"note","title":"Autonomous Driving Reality Check","text":"Despite hype, self-driving cars are not fully deployed. Most systems are Level 2 (driver assistance) or Level 3 (conditional automation). Full autonomy (Level 5) remains a research challenge."},"narration":"AI is already part of your daily life: from email filters to generative AI. But remember, these are all narrow AI — they do one thing well."},{"kind":"content","heading":"Step-by-Step: Building a Simple Rule-Based System","body":"Let's create a diagnostic system for plant health using explicit rules. This is not machine learning — it's a classic AI approach.\n\n1. Define the problem: Diagnose plant health from symptoms.\n2. Gather data (symptoms): Leaf color (Green, Yellow, Brown) and leaf texture (Healthy, Spots, Wilting).\n3. Define rules:\n - IF leaf color = Yellow AND texture = Healthy → Nutrient Deficiency\n - IF leaf color = Brown AND texture = Wilting → Lack of Water\n - IF leaf color = Green AND texture = Spots → Fungal Infection\n - IF leaf color = Green AND texture = Healthy → Healthy\n4. Implement: Code these as IF-THEN statements.\n5. Test: Input (Yellow, Healthy) → Output: Nutrient Deficiency.\n\nThis illustrates how AI can encode human expertise. Modern ML learns such rules from data automatically.","callout":{"variant":"exercise","title":"Try It Yourself","text":"Think of a simple decision you make daily (e.g., what to wear based on weather). Write 3-4 rules. That's a tiny AI system!"},"narration":"Even simple rule-based systems are AI. This example shows how explicit rules can make decisions. Machine learning automates rule discovery."},{"kind":"quiz","heading":"Check Your Understanding","questions":[{"question":"Which type of AI is currently in use today?","options":["Narrow AI","General AI","Super AI","All of the above"],"questionId":"cmrf73k280000pd27g6as6fbc"},{"question":"What is the correct hierarchy from broadest to most specific?","options":["AI → Deep Learning → Machine Learning","AI → Machine Learning → Deep Learning","Machine Learning → AI → Deep Learning","Deep Learning → Machine Learning → AI"],"questionId":"cmrf73k280001pd27brnfku95"},{"question":"Which component is considered the 'lifeblood' of a machine learning model?","options":["Algorithm","Data","Model","Evaluation"],"questionId":"cmrf73k280002pd27dcbr36xb"}],"quizId":"qz_cmk5q6sfo0004g4eogroxjp9r"},{"kind":"summary","heading":"Lesson Summary","takeaways":["AI is about machines performing tasks that require human intelligence.","Only narrow AI exists today; general and super AI are hypothetical.","AI is the broad field; ML learns from data; DL uses deep neural networks.","Key ML components: data, algorithms, model, training, evaluation.","Real-world examples include spam filters, generative AI, and medical diagnosis.","Rule-based systems are a simple form of AI; ML learns rules from data."],"narration":"You now have a solid foundation. AI is not magic — it's a tool. In the next lesson, we'll dive into how machine learning actually works."}]}