What Is Intelligence, Anyway?
Can a machine actually "think," or is it just really good pattern matching?
Here's a trick question: is your thermostat intelligent? It senses the temperature, compares it to a number you set, and flips the furnace on or off. That's a real decision based on real data. But nobody calls a thermostat "smart" the way we'd call a person smart. So what's missing?
That gap is exactly where this course lives. Most of what gets marketed as "AI" today—from the autocomplete on your phone to the recommendation feed deciding what video plays next—is what researchers call narrow AI: systems that are extremely good at one specific job and clueless about everything else. A chess engine that can beat any human on Earth has no idea what chess even is outside of the board in front of it. It can't order a pizza. A self-driving car that can navigate a highway better than you can has zero opinion about anything that isn't a road.
Compare that to general AI (AGI)—a system that could learn anything a human can, the way a person can go from cooking dinner to debating politics to fixing a bike without being "retrained" for each one. AGI doesn't exist yet, despite how often it shows up in headlines. Everything you interact with today, no matter how impressive, is narrow.
So how do we even decide what counts as "thinking"? In 1950, mathematician Alan Turing dodged the philosophical mess of defining intelligence and proposed a test instead: if a human chatting with a hidden computer can't reliably tell it apart from a hidden human, the machine passes. It's a clever sidestep—it judges behavior, not whether something "really" understands. That distinction (acting intelligent vs. actually understanding) is one you'll keep bumping into all course long, especially once we get to chatbots that sound thoughtful but are doing something very different under the hood.
By the end of today, you won't have a clean, one-sentence definition of intelligence—nobody does, including the experts. But you'll have a sharper set of questions to ask whenever someone says "this is AI," which turns out to be a far more useful skill.
Try It Yourself
Is It AI?
For each example, decide whether it counts as AI and pick the explanation that best matches how it actually works.
1. A thermostat that turns the heat on when the temperature drops below 68°F.
2. A chess engine that has beaten every human grandmaster.
3. A voice assistant that can set a timer, but also fails badly on an unfamiliar accent.
4. A calculator that computes 847 × 392 instantly.
Want to go deeper?
What Is Artificial Intelligence? — Crash Course AI #1
For Teachers: Full Lesson Plan Detail
- Define intelligence using multiple competing definitions
- Distinguish between narrow AI, general AI, and superintelligence
- Evaluate real-world AI examples against a working definition of intelligence
1. Warm-Up
"Is It AI?" sorting activity—students sort examples (a thermostat, a chess engine, Siri, a calculator, a self-driving car) into AI / not AI and defend their choices.
2. Direct Instruction
Mini-lecture on the spectrum from rule-based automation to true learning systems; introduce the Turing Test and its limitations.
3. Guided Practice
Small groups design a 5-minute "Turing Test" style game to try to distinguish a human from a bot.
4. Discussion
Journal response: "Do you think today's chatbots pass the Turing Test? Why does it matter if they do?"
Assessment: Exit ticket: one paragraph defining AI in the student's own words with one supporting example.