2.3. Reflection Checkpoint: AI Workloads and Responsible AI Mastery

What breaks if you skip this checkpoint? You'll move forward without confirming the fundamentals, and gaps in understanding will compound. Imagine reaching Phase 6 and realizing you still confuse Computer Vision with NLP—every subsequent concept builds on these foundations.

Consider each question below like a diagnostic test. For instance, if you hesitate on "Is handwriting extraction NLP or Computer Vision?"—that hesitation signals a gap. Review Section 2.1 before proceeding. These aren't trick questions; they're validation that your mental models are solid.

1. An application analyzes photos to identify products on store shelves. Which AI workload category is this?

   • Computer Vision. The input is images (photos), so apply the Input Modality Framework from Section 1.2.1.

2. A company extracts key phrases from customer reviews to identify trending topics. Which workload is this?

   • NLP (Natural Language Processing). The input is text (reviews).

3. A bank is building an AI system to approve loans. Which Responsible AI principle is most important for avoiding bias against protected groups?

   • Fairness. Recall that fairness ensures AI doesn't discriminate based on gender, ethnicity, or other protected characteristics.

4. Which Responsible AI principle ensures users understand they are interacting with an AI system?

   • Transparency. This principle requires users know the system's purpose, limitations, and that it's AI-powered.

5. A user uploads a handwritten note and wants to convert it to typed text. Is this NLP or Computer Vision?

   • Computer Vision (OCR). The INPUT is an image, even though the output is text. Always classify by input type.