ai 과정, 앱개발의 방향성

Main ideas / lessons conveyed

• Purpose of the class (AI-based digital learning environment design)

  • Learn how to design lessons using AI tools and how to use EdTech tools effectively.
  • Connect this to pedagogy frameworks (e.g., T-Pack, A-Pack).
  • Prepare trainees for their future role as teachers who must evaluate and support learners using AI.

• Caution: AI and EdTech are not automatically “good”

  • AI/EdTech can be beneficial, but the key is not “use tools” vs “don’t use tools.”
  • The reason to use tools is to better achieve educational goals—especially by collecting individual learning data to support:
    • learner analysis
    • class preparation

• Real concern: assessment/evaluation is breaking under generative AI

  • AI cheating is being detected even at top Korean universities, but detectors lag behind because AI performance improves rapidly.
  • Education must rethink evaluation because it’s becoming hard to tell whether students:
    • received light help from AI, or
    • outsourced their thinking to AI.

• Redefine “human competencies” and the essence of learning in the AI era

  • Evaluation should not only check achievement or whether AI was used.
  • Education needs to clarify:
    • which competencies matter,
    • what learning is aiming for, and
    • how those competencies can be evidenced when AI changes the production process.

• Core evaluation shift: focus on cognition and learning processes

  • Generative AI can summarize, write, code, etc., acting like a “cognitive partner.”
  • The question is not only “Was AI used?”
  • It is: What is happening inside the learner’s cognitive structure?
    • Is AI replacing student thinking?
    • Is it expanding thinking?
    • Is it reconstructing thinking?

• How to reflect while doing assignments

  • As students work (including in other classes), they should ask whether they are:
    • outsourcing thought to AI,
    • using AI only as a reference while leading themselves, or
    • having their perspective expanded by what AI reveals.

• Design assessment according to lesson objectives

  • If the goal is basic thinking / conceptual understanding, assessments should reveal internal cognition, which may require:
    • regulating/monitoring AI use
    • evaluating whether learning occurred properly
  • If the goal is real-world problem solving and AI is used to improve output quality, evaluation should include:
    • AI utilization capability (how effectively and appropriately AI is used)
  • Mentioned assessment approaches:
    • metacognitive / critical analysis of AI-generated answers (already used by some professors)
    • more performance-based assessments requiring explanation of the problem-solving process
    • subject areas that naturally expose cognition (e.g., math/physics/chemistry)

---

Detailed methodology / instructions for app + assessment direction

A) When planning AI learning assessment (process-focused design)

• Start from lesson objectives, not from “detect cheating.”
• Decide what the assessment should center on:
  • learner thinking/cognition, or
  • performance produced with an AI system
• Consider whether you need AI governance:
  • Use/limit AI access if you must measure internal understanding.
• Select the assessment type accordingly:
  • Metacognitive evaluation: learners critique/reflect on AI outputs.
  • Process-based performance assessment: learners describe how they solved the problem.
  • AI utilization criteria: if AI is intended to be used for real tasks, explicitly evaluate AI use quality.

B) When developing an AI teaching app (set direction using “Effect With vs Effect Of”)

• Use the tool-effect distinction to decide the app’s intended learning impact:
  • Effect with (tool-supported performance improvement)
    • Performance improves while using the tool.
    • The ability disappears when the tool is removed.
  • Effect of (cognitive restructuring / internalization)
    • Using the tool restructures learner cognition.
    • The benefit remains even after the tool is removed.
• Choose an app direction aligned with your objectives:
  • Do you want a tool that improves results only with AI?
  • Or a tool that also develops competencies independently of AI later?

C) Choose AI’s role in the classroom (replace / scaffold / partner / feedback)

• AI roles described:
  • Replacement
    • AI performs the task; the student edits/evaluates.
    • Best suited when the target is information acquisition.
  • Scaffolding
    • AI asks questions and provides hints so the student is the main actor in thinking.
    • Goal: internalize the thinking process.
  • Partner
    • AI simulates discussion/practice so students get authentic near-field practice.
    • Example: an AI interview system for interview practice.
  • Feedback
    • AI reacts to student output so students revise and reflect.
    • Reflection must be intentionally designed.

D) Development practicality advice (for the course assignment)

• Since the assignment is to develop an AI application:
  • Clarify app goals and AI role first (based on the frameworks above).
  • Start early; small daily progress + AI assistance helps.
  • Demonstrate that an app can be produced quickly when using an AI model.
• Workflow/tooling approach:
  • “vibe coding” (rapid coding with AI), plus the need for debugging
  • Misconception corrected: even if AI can code, you still need Vibe Debugging because AI may not meet your required quality/style.
• Recommended practical approach:
  • the “Loo-up method” (described) to create desired results with less friction, while still debugging.
• Credit limitations:
  • even with limited daily credits (e.g., 5/10/15 mentioned), development is still possible if started early.

E) Gamification as an app design strategy (with caution)

• Gamification definition:
  • maximize user experience and participation using game mechanics/rewards
  • build voluntary motivation through fun and immersion
• Key caution:
  • if students become too immersed in the game, educational goals may be missed
  • the instructor’s design determines whether goals are achieved

---

Example app described (for inspiration)

• The instructor demonstrated an AI teaching tool: “Vibe Coding Encyclopedia”
• Features:
  • Game-like structure using packs, cards, ranks, and coins
  • Students earn mastery titles/coins based on quiz performance timing and hint usage
  • Content progression such as basic → practical → advanced → expert → rare/legendary
  • A time limit per level, making “ask AI for everything” less effective
  • System error fixing when terms don’t appear correctly
  • Pokedex/storage-style view of learned terms
  • Challenge mode and review prompts
  • Classroom/ranking elements to encourage competition
  • An administrator page for managing classes
• Intended learning effect:
  • encourage learning coding-related terms in advance, sometimes even before teacher explanation
  • increase engagement and effort through competition and fun

---

Closing focus / takeaway

• The class (though 1 credit) emphasizes helping students make an impact in the AI era through thoughtful app/assessment design.
• Students are encouraged to:
  • start designing early,
  • adapt ideas to their own contexts, and
  • ensure plans align with educational goals and evaluation methods.

---

Speakers / sources featured

• Hwang Hyun-woo (primary instructor throughout)
• Examples mentioned:
  • Korea University, Seoul National University, Yonsei University (where AI cheating detection occurred)
• Direction referenced:
  • South Korea’s Ministry of Education (increasing AI tool use for digital literacy, AI utilization, and problem-solving skills)
• General references:
  • unnamed scholars/professors (for educational evaluation concepts and metacognitive assessment approaches)