Course Overview
AI in Action – Level 4 advances students from “understanding how AI works” (Level 3) to designing, improving, and deploying AI solutions with stronger technical thinking and responsible decision-making. Students will move beyond recognizing AI concepts to building smarter pipelines (data → model → evaluation → deployment), comparing approaches (rule-based vs learning-based), and practicing real-world skills such as dataset design, bias detection, model testing, iteration, and impact evaluation.
Throughout the course, students will create practical AI projects using computer vision, NLP, and robotics-style interaction, while learning how to keep AI fair, safe, private, and useful.
Course Details
In Level 4, students:
- Study deployment, privacy, and impact through realistic school/community case studies and responsible AI practices.
- Refine their understanding of the AI process by investigating model behaviours, errors, and limitations.
- Develop stronger data literacy: collecting better data, cleaning it, labelling it, balancing it, and validating it.
- Learn to evaluate models professionally (not just “it works”): accuracy vs real-world reliability, confusion matrices (simplified), and testing with unseen examples.
- Explore ML types more deeply, including how to choose the right approach for a task.
- Build richer human-computer interaction experiences (text + vision + UI decisions), and learn how design affects trust and usability.
Course Objectives
By the end of this course, students will be able to:
AI Concepts & Process
- Explain why AI models make mistakes and how to reduce errors through iteration.
- Describe the AI lifecycle at a deeper level: problem → data → training → testing → improvement → deployment → monitoring.
- Compare solutions and justify choices (e.g., rule-based vs ML, or supervised vs unsupervised).
Data Skills
- Plan and collect a dataset with clear labels, balance, and quality checks.
- Identify and fix common data problems: missing values, noisy data, unbalanced classes, biased sampling.
- Represent and visualize data more thoughtfully to support decisions.
Machine Learning Skills
- Train a simple model and evaluate performance using clear testing routines (including “new/unseen” cases).
- Interpret results using student-friendly evaluation tools (e.g., error counts, class-by-class performance, simplified confusion view).
- Improve a model by adjusting data, labels, and testing strategy.
Human-Computer Interaction
- Design an AI interface that is clear, ethical, and user-friendly (prompts, feedback, confidence messages, error handling).
- Explain how UI choices can increase/decrease trust in AI systems.
Ethics, Privacy, and Impact
- Present an AI solution’s impact with benefits, risks, and safeguards.
- Apply privacy and consent principles (what should/shouldn’t be collected, stored, or shared).
- Identify fairness risks and propose practical mitigation steps.