2.1.1. Generative AI vs. Traditional AI
💡 First Principle: Choose generative AI when you need content creation, summarization, or natural language interaction. Choose traditional AI when you need classification, prediction, or pattern detection. This decision tree prevents the common mistake of using the wrong tool for the job—like using a creative writer to sort mail, or a sorting machine to write poetry. The business outcome determines the AI type.
The following comparison helps you match business needs to AI approaches:
| Characteristic | Generative AI | Traditional AI |
|---|---|---|
| Primary function | Creates new content | Classifies/predicts from patterns |
| Output | Text, images, code, summaries | Categories, scores, predictions |
| Training approach | Large language models, massive datasets | Task-specific training data |
| Example use cases | Draft emails, summarize meetings, answer questions | Fraud detection, demand forecasting, image recognition |
| Cost model | Token-based (input + output) | Often flat or compute-based |
⚠️ Exam Trap: The exam may describe a prediction or classification task and offer generative AI as an option. Remember: if the business needs to categorize or predict, traditional AI is usually more appropriate.
Reflection Question: A bank wants to detect fraudulent transactions in real-time. Should they use generative AI or traditional AI? Why?
