Module 1: What is AI?
Lesson Overview
In this lesson, we explore the three main categories of Artificial Intelligence: Narrow AI, General AI, and Superintelligent AI. Each category reflects a different level of capability and ambition — from practical tools we use today, to theoretical systems that could one day surpass human intelligence. Understanding these categories helps you grasp where AI is now, and what the future may hold.
1. Narrow AI (Also Called Weak AI)
Definition:
Narrow AI refers to AI systems that are designed to perform a specific task, or a limited range of tasks, extremely well. These systems are trained on data and programmed to operate within strict boundaries.
Examples:
Voice assistants like Siri or Alexa
Spam filters for email
Facial recognition on smartphones
Recommendation engines on Netflix or Spotify
Navigation systems in Google Maps or Waze
Chatbots in customer service
Key Features:
Task-specific intelligence
Cannot operate outside its programmed domain
No self-awareness or understanding
Commonplace and widely deployed today
Narrow AI might appear intelligent, but it doesn't "understand" what it's doing. A chatbot that helps you book a flight doesn't actually know what a flight is — it recognises patterns in language and responds accordingly.
Narrow AI is where we are now. Nearly all AI applications currently in use — from translation tools to predictive text — fall into this category.
2. General AI (Also Called Strong AI or AGI)
Definition:
General AI refers to a type of AI that can perform any intellectual task that a human being can do. It would be capable of learning, reasoning, understanding context, adapting to new situations, and transferring knowledge from one task to another.
Key Features:
Human-level problem-solving abilities
Learns new tasks without retraining from scratch
Can apply reasoning across multiple domains
Thinks abstractly, understands context and nuance
Not yet achieved — purely theoretical at this stage
If Narrow AI is like a chess master who can only play chess, General AI is like a human who can learn to play chess, write a poem, do algebra, and then go cook dinner — all without being reprogrammed.
No current AI system — not even ChatGPT — meets the definition of General AI. While tools like ChatGPT seem flexible, they are still confined to pattern recognition and lack consciousness or real understanding.
3. Superintelligent AI
Definition:
Superintelligent AI refers to a hypothetical future form of AI that surpasses human intelligence in every measurable way — creatively, socially, emotionally, and intellectually.
Key Characteristics:
Smarter than the best human minds across all disciplines
Capable of self-improvement
Could solve problems humans cannot even comprehend
Raises major ethical and existential questions
Superintelligence is the subject of both fascination and concern. Thinkers like Nick Bostrom and Elon Musk have warned that once AI reaches superintelligence, it could evolve beyond our control — potentially making decisions that affect humanity’s survival.
This isn’t science fiction anymore. Though we are nowhere near creating such a system, rapid progress in AI research means these discussions are increasingly important — particularly around safety, ethics, and governance.
Comparing the Categories
Here’s a simple table to summarise the key differences:
CategoryCurrent StatusTask ScopeIntelligence LevelNarrow AIWidely usedOne specific taskBelow humanGeneral AI (AGI)Not yet developedAny human-level intellectual taskEqual to human (in theory)Superintelligent AIHypotheticalAny task, far beyond human limitsGreater than human (in all areas)
Why It Matters to Understand These Categories
Knowing the differences between types of AI helps you:
Set realistic expectations about what AI can and can’t do today
Understand where we’re heading, and how fast
Separate hype from reality, especially in media or marketing
Think critically about risks and regulation, especially when it comes to general or superintelligent AI
Right now, we’re building and using narrow AI — systems that are useful, but ultimately limited. General AI is the next big milestone, and while it might take decades to develop, it will change everything when it arrives. Superintelligent AI may sound far off, but it’s a real possibility that’s taken very seriously by experts in AI safety and ethics.
Quick Summary
Narrow AI performs specific tasks and is common today.
General AI would match human intelligence across all areas — but doesn’t yet exist.
Superintelligent AI would surpass humans entirely, raising major ethical concerns.
Understanding these categories helps us engage responsibly with AI today — and think wisely about its future.
Reflection Question
If you could design a General AI to help with one area of your life or work, what would it be? What tasks would you trust it to do — and where would you want to stay in control?