Narrow AI
AI that does ONE thing very well. ChatGPT for writing, image generators for images, chess engines for chess. All the AI you use today is narrow AI.
This is what you'll encounter in real life.
Step 1 of 3
Understand the basics.
Start with simple definitions. Learn what AI is good at and what it struggles with—without the computer science textbook.
Definition
What is AI, really?
AI stands for Artificial Intelligence. In simple terms, it's software that can learn from examples and make decisions without being explicitly programmed for every single situation.
Think of it like teaching a child. You don't have to explain every single scenario—they learn patterns and apply them to new situations. AI works similarly: it learns from lots of examples and recognizes patterns.
Categories
The three types of AI (simplified)
AI comes in different levels of sophistication. Understanding the differences helps you know what to expect.
General AI
AI that can do MANY things like a human brain. It doesn't exist yet. This is theoretical—AI that could learn any task, like humans can.
Scientists are still working on this.
Super AI
AI that surpasses human intelligence in every way. This is science fiction right now. Don't worry about it—we're nowhere close.
Keep it in perspective.
The Simple Version
How does AI actually work?
Here's the non-technical explanation:
1. Training
AI learns from millions of examples. If you're training an image recognizer, you show it thousands of cat photos. It notices patterns: pointy ears, whiskers, fur texture.
2. Pattern Recognition
The AI creates an internal "rule book" (called a neural network). It's not programmed by humans—it learns these rules automatically from the examples.
3. Prediction
When you show it a new cat photo, the AI applies those learned patterns to guess: "This looks like a cat because it has these features." That's it. No magic.
The key insight: AI is just very sophisticated pattern matching. It's not magic, and it's not thinking. It's finding patterns in data, then using those patterns to make predictions.
Strengths
What AI excels at
- Processing large amounts of text Summarizing documents, writing emails, answering questions based on information it learned.
- Finding patterns in images Recognizing objects, generating images, reading handwriting, identifying faces.
- Speed and consistency Analyzing data 1,000x faster than humans. No fatigue, no inconsistent judgment.
- Creative tasks with constraints Brainstorming ideas, generating code, writing marketing copy—within guardrails.
Limitations
What AI struggles with
- Common sense AI can't understand context the way humans do. "Light" can mean illumination or "not heavy"—it gets confused.
- Explaining itself AI can give you an answer, but it can't always explain WHY. It's a black box sometimes.
- Handling the unexpected If training data doesn't include something, AI struggles. Show it a cat photo in a weird angle it's never seen, it might fail.
- Making predictions far into the future AI is good at next-week predictions, not great at next-year. The further out, the less reliable.
- Understanding human values AI doesn't inherently know what's "fair" or "ethical." These are human judgments we have to teach it.
Real World
AI in everyday life
- Your email spam filter
- Trained on millions of emails to recognize patterns of spam. Learns your personal preferences over time.
- Netflix recommendations
- Analyzes what you and millions of others watch, finding patterns in taste. Then suggests movies that similar people liked.
- Your phone's autocorrect
- Learns from millions of texts to predict the next word you'll type. Gets better the more you use it.
- ChatGPT & Claude
- Trained on billions of words to predict the next word in a sentence. Chat with it naturally, and it generates responses word-by-word.
- Google Translate
- Learned translation patterns from millions of translated documents. Not perfect, but surprisingly useful.
- Medical imaging diagnosis
- Trained to recognize cancer, fractures, and anomalies in X-rays. Works alongside doctors, not replacing them.
Key takeaways
1. AI is pattern matching
Not magic. Not conscious. Just math recognizing patterns in data.
2. Today's AI is "narrow"
Good at one thing. ChatGPT writes; image generators generate images. That's it.
3. AI has real strengths and real limits
It's fast and consistent but lacks common sense. Use it as a tool, not as truth.
4. You already use AI
In spam filters, recommendations, autocomplete. It's everywhere, usually invisible.
Ready for step 2?
Now that you understand the basics, let's get hands-on. Try writing a prompt and see how AI responds.
Go to Step 2: Try a Prompt