Prismo | Branches of AI (Part 1)
Machine Learning & Logic-Based AI: Two Roads to “Smart” Machines
When people talk about Artificial Intelligence, they often lump everything together under one mysterious umbrella. But AI isn’t just one thing — it’s a collection of different approaches. Today, we’ll look at two of the earliest and most important “branches” of AI: Machine Learning and Logic-Based AI.
AI in Plain Language
Think of AI like a toolbox. Inside are different “tools” that engineers use to make machines act smart. Two of the most common tools are:
Machine Learning (ML): Instead of writing out all the rules, you feed the machine lots of data and let it learn patterns.
Logic-Based AI: Instead of learning, this branch relies on rules and logic written by humans, like an elaborate decision tree.
Both aim to make machines act intelligently, but they get there in very different ways.
Why It Matters for You
Machine Learning is everywhere today: fraud detection, personalized ads, recommendation systems, even predicting flight delays.
Logic-Based AI may sound old-fashioned, but it still powers rule-heavy areas like compliance checks, medical diagnosis systems, and expert systems.
Knowing the difference helps you understand why some AI tools improve over time (learning), while others stay rigid but dependable (rule-based).
Real-World Examples
Machine Learning:
Netflix recommendations that improve as you watch more.
Gmail’s smart reply suggestions.
Banks using ML to flag unusual spending.
Logic-Based AI:
Tax preparation software applying thousands of rules.
Early medical expert systems like MYCIN (1970s) that diagnosed infections.
Chatbots with scripted “if this, then that” pathways.
A Few Limitations to Keep in Mind
Machine Learning: Needs huge amounts of data and can be a “black box” (hard to explain how it reached a decision).
Logic-Based AI: Struggles when problems get messy — you can’t write rules for every possible situation.
Try It Yourself
Choose any AI app on your phone that supports image recognition — like ChatGPT (with vision), Claude, or Gemini.
1. Take a picture of an everyday object (for example: mug, shoe, chair).
2. Ask the AI: “What’s in this picture? Describe it.”
3. After you get the answer, ask the AI again: “How do you know it’s that object?” The AI will explain using patterns and features (shapes, textures, context).
4. Now try to write 5 exact rules that always define that object. For example:
If it has a handle → mug (but what about bowls with handles?)
If it’s cylinder-shaped → mug (but what about vases?)
You’ll quickly see that rule-based definitions break down, while ML handles variety with ease. That’s the real difference between the two branches.
Think About It
In your daily work, where do you see systems that are clearly rule-based? Where do you notice systems that seem to learn from experience?
The Takeaway
AI isn’t one-size-fits-all, and these two branches show just how different approaches can be. Machine Learning thrives on data and improves over time, while Logic-Based AI depends on human-crafted rules and shines in areas where consistency is key.
At Prismo: The 10-Minute AI School, we’ll continue breaking down these branches so you can see how they fit into your world of work.
💡 Curious about how this fits into your own career? Drop us a note — we love a good AI chat.
