1.1.1. Defining Artificial Intelligence
💡 First Principle: AI is software that performs tasks associated with human cognition — perceiving (vision, speech), reasoning (drawing conclusions), and deciding (choosing actions). The defining trait isn't how it works internally, but what it accomplishes: behavior that would look intelligent if a person did it.
A spam filter that flags junk mail, a navigation app that picks the fastest route, a chatbot that answers questions — all are AI, even though they work in completely different ways internally. What unites them is that each handles a task we'd otherwise need a human to do. This is why "AI" alone is rarely a precise enough answer; the exam usually wants you to identify the specific kind of AI workload involved (vision, language, speech, generative, and so on), which Phase 3 covers in detail.
⚠️ Exam Trap: Don't assume "AI" always means a modern neural network. A traffic light controller following fixed if-then rules can be a form of AI, but it does no learning. When a question contrasts "AI" with "machine learning," the distinction being tested is usually that ML systems improve from data while plain AI may just follow hard-coded logic.
Reflection Question: A company hard-codes a set of business rules to approve or deny loan applications. Is this AI? Is it machine learning? Why might the answer differ for each?
