--- id: wiki-2026-0508-ai-literacy title: AI Literacy category: 10_Wiki/Topics status: verified canonical_id: self aliases: [AI λ¬Έν•΄λ ₯, AI fluency, AI competence, co-intelligence, AI education] duplicate_of: none source_trust_level: B confidence_score: 0.85 verification_status: conceptual tags: [ai-literacy, education, prompt-engineering, critical-thinking, ai-ethics, future-skills, productivity] raw_sources: [] last_reinforced: 2026-05-09 github_commit: pending inferred_by: Claude Opus 4.7 (manual cleanup 2026-05-09) tech_stack: language: education / process applicable_to: [Education, HR, Self-development] --- # AI Literacy ## πŸ“Œ ν•œ 쀄 톡찰 (The Karpathy Summary) > **AI μ‚¬μš© β‰  AI literacy**. λ§€ user κ°€ ChatGPT μ‚¬μš© κ°€λŠ₯ κ°€, "μ™œ 이 λ‹΅?", "이 λ‹΅ κ°€ μ‹ λ’°?", "μ–΄λ–»κ²Œ 더 잘?" 의 λ‹΅ κ°€ 별 skill. **Understand + Utilize + Evaluate + Reflect** 의 4 dimension. μ½”λ”© literacy 의 modern ν›„κ³„μž. ## πŸ“– κ΅¬μ‘°ν™”λœ 지식 (Synthesized Content) ### 4 dimension (Long et al. 2020 + Mollick "Co-intelligence") 1. **Understand**: AI / LLM 의 μž‘λ™ 원리. - Token / context window / hallucination. - Train data 의 cutoff. - λ§€ model 의 strength / weakness. - Pattern matching β‰  reasoning. 2. **Utilize**: 효과적 μ‚¬μš©. - Prompt engineering (clear task, examples, constraints). - Tool / agent (Claude Code, Cursor). - Multi-step task 의 break down. - λ§€ model 의 적절 선택 (Opus κ°€ critical, Haiku κ°€ quick). 3. **Evaluate**: κ²°κ³Ό 의 verification. - Hallucination detect (citation, fact-check). - Bias 인지 (training data 의 bias). - Source verification. - 자기 reasoning 의 sanity check. 4. **Reflect**: 윀리 / μ‚¬νšŒ 의 영ν–₯. - Privacy (λ§€ prompt κ°€ server). - Bias 의 amplification. - Job displacement. - Misinformation potential. - 의쑴 / μ˜μ§€ 의 μœ„ν—˜. ### Bloom's Taxonomy 식 (AI 의) 1. **Remember**: terminology (LLM, RAG, fine-tune). 2. **Understand**: concept (token, context, hallucination). 3. **Apply**: use ChatGPT / Claude 의 task. 4. **Analyze**: λ§€ output 의 quality / bias 의 critique. 5. **Evaluate**: μƒˆ model / tool 의 verdict. 6. **Create**: prompt template / agent / λ§€ workflow. β†’ λ§€ level 의 progressive skill. ### λ§€ audience 의 priority #### General public - AI κ°€ 무엇 (LLM, image gen). - ChatGPT / Claude μ‚¬μš©. - Hallucination 인지. - Privacy + scam 인지. #### Professional (non-tech) - λ§€ task 의 AI ν™œμš© (writing, analysis). - Prompt 의 효과적 μž‘μ„±. - Output 의 critical review. - Tool 의 적절 (research, coding, design). #### Engineer - API integration. - RAG / fine-tune / agent. - Cost / latency / quality optimization. - Production eval. #### Educator / parent - Cheating 의 detect (hard). - AI-augmented learning (Khan Academy). - 학생 의 critical thinking. #### Policymaker - Regulation (EU AI Act, US EO). - Bias / fairness. - Job market impact. ### Common misconception 의 fix - ❌ "AI κ°€ 만물 μ•Œμ•„" β†’ μ‹€μ œ = train cutoff + hallucination. - ❌ "AI κ°€ sentient" β†’ pattern matching. - ❌ "AI = AGI κ³§" β†’ λ§€ capability 의 plateau / breakthrough cycle. - ❌ "AI κ°€ magic" β†’ train data + μ•Œκ³ λ¦¬μ¦˜. - ❌ "Prompt κ°€ κ³ μ •" β†’ iterate + variant. - ❌ "μ΅œμ‹  model κ°€ 항상 best" β†’ cost / latency trade-off. ### λ§€ ν‰μ†Œ work 의 AI integration #### Writing - Draft β†’ AI revise β†’ human polish. - λ§€ paragraph 의 alternative. - Tone / style 의 adjust. #### Coding - Boilerplate generation. - Bug fix. - Refactor. - Test write. #### Research - Literature 의 summarize. - Brainstorm. - λ§€ paper 의 critical question. #### Decision - Pros / cons matrix. - λ§€ option 의 risk. - Analogous case 의 search. β†’ λ§€ task 의 적절 use case. ### Critical thinking β€” λ§€ AI output 의 fact-check 1. **Source**: AI κ°€ source claim? 2. **Reproducibility**: λ‹€λ₯Έ model 도 같은 λ‹΅? 3. **Plausibility**: λ§€ number / claim 의 sanity. 4. **Bias check**: λ§€ perspective. 5. **Counter-question**: "μ™œ μ•ˆ X?". ## πŸ’» νŒ¨ν„΄ (ν™œμš© + 검증) ### Prompt 의 4 component ``` 1. Role / context: "You are an expert React dev." 2. Task: "Refactor this code to use hooks." 3. Constraints: "Keep TypeScript types. No new dependencies." 4. Format: "Output: code block + 1 sentence summary." ``` ### Iteration pattern ``` v1: "Translate to French: Hello" v2: "Translate to formal French: Hello" v3: "Translate to formal French (business email register): Hello" β†’ λ§€ iter 의 specificity ↑. ``` ### Few-shot example ``` "Classify sentiment. Example 1: 'This is great!' β†’ positive Example 2: 'I hate it' β†’ negative Example 3: 'Mediocre' β†’ neutral Now: '{user_input}' β†’ ?" ``` β†’ λ§€ task 의 일관 output. ### Chain of thought ``` "Solve step-by-step. Q: A train travels 60 mph for 2.5 hours. Distance? Reasoning: ..." ``` β†’ Math / logic 의 μ •ν™• ↑. ### Self-verification ```ts async function answer(query: string) { const ans1 = await llm.complete(query, { temp: 0.7 }); const ans2 = await llm.complete(query, { temp: 0.7 }); if (similar(ans1, ans2)) return ans1; // λ‹€λ₯Έ λ‹΅ = μ˜μ‹¬. const verify = await llm.complete(`Q: ${query}\nA1: ${ans1}\nA2: ${ans2}\n\nWhich more accurate? Why?`); return verify; } ``` ### Hallucination detect ```ts function fact_check(claim: string, search_result: string) { return llm.complete(`Claim: ${claim}\nSource: ${search_result}\n\nDoes source support claim? (Y/N + reason)`); } ``` ### Multi-step task ``` "Build a flashcard app": 1. Define data structure. 2. Choose framework (React). 3. Component breakdown. 4. State management. 5. Persistence. 6. Test plan. β†’ λ§€ step 의 separate prompt. ``` ### Curriculum (자기 ν•™μŠ΅) ``` Week 1: ChatGPT basic + 맀일 1 task. Week 2: Prompt engineering deep. Week 3: Claude + Cursor 의 λ‹€μ–‘ tool. Week 4: API call (Python). Week 5: RAG basic. Week 6: Agent basic. β†’ 6 week 의 baseline literacy. ``` ## πŸ€” μ˜μ‚¬κ²°μ • κΈ°μ€€ (Decision Criteria) | 상황 | AI μ‚¬μš© | 검증 | |---|---|---| | Brainstorm | βœ… Generate ideas | Filter manually | | Draft writing | βœ… Initial + iterate | Polish + fact-check | | Code | βœ… Boilerplate, bug | Test + review | | Research | βœ… Summary, search | Cite + verify | | Critical decision | βœ… Pros/cons | λ§€ source 의 check | | Sensitive (legal, medical) | ⚠️ Reference only | μ „λ¬Έκ°€ + final | | Personal expression | ❌ Authentic κ°€ κ°€μΉ˜ | | | μ‹œν—˜ / 평가 | λ§€ institution 의 policy | | **κΈ°λ³Έκ°’**: AI = λ§€ task 의 첫 draft / brainstorm. Human = final review + critical decision. ## ⚠️ λͺ¨μˆœ 및 μ—…λ°μ΄νŠΈ (Contradictions & Updates) - **"AI 의 이해" 의 standard**: λ§€ country / institution 의 λ‹€λ₯Έ curriculum. - **Tool change 의 빠름**: 6 month λ§ˆλ‹€ update. "μ΅œμ‹  best practice" κ°€ short-lived. - **Critical thinking 의 paradox**: AI 의 λ‹΅ 의 검증 = AI μ‚¬μš©. Meta-cognition ν•„μš”. - **Education 의 cheating**: λ§€ 학생 의 ChatGPT μ‚¬μš© = 곡정 X / μƒˆ reality? - **AI literacy β‰  AI fear**: λ§€ risk 의 인지 의 healthy. - **Co-intelligence (Mollick)**: λ§€ task 의 AI + human 의 collaboration. "AI κ°€ μ—†λŠ” 인간 의 의미" μž¬μ •μ˜. ## πŸ”— 지식 μ—°κ²° (Graph) - λΆ€λͺ¨: [[Digital-Literacy]] Β· [[Information-Literacy]] - λ³€ν˜•: [[Prompt_Engineering|Prompt-Engineering]] Β· [[AI-Ethics]] - Adjacent: [[Lifelong-Learning]] - μ‘μš©: [[Code Agent β€” Devin / Cursor / Claude Code]] ## πŸ€– LLM ν™œμš© 힌트 (How to Use This Knowledge) **μ–Έμ œ 이 지식을 μ“°λŠ”κ°€:** - λ§€ team 의 AI training program. - 학ꡐ / νšŒμ‚¬ 의 AI policy μž‘μ„±. - λ§€ user 의 self-development plan. - AI tool 의 onboarding. - "AI κ°€ 일자리 λΉΌμ•—?" 질문 의 nuanced λ‹΅. **μ–Έμ œ μ“°λ©΄ μ•ˆ λ˜λŠ”κ°€:** - AI 의 specific 기술 detail (λ‹€λ₯Έ doc). - Specific νšŒμ‚¬ 의 AI training (custom curriculum). - 맀우 κ³ κΈ‰ (researcher 의 already known). - AI 의 banning / blanket prohibition (λ‹€λ₯Έ framework). ## ❌ μ•ˆν‹°νŒ¨ν„΄ (Anti-Patterns) - **"AI κ°€ λͺ¨λ“  κ±° λ‹΅": critical 검증 X. - **AI 의 output 의 100% trust**: hallucination 의 victim. - **Privacy λ¬΄μ‹œ**: sensitive prompt κ°€ server. - **Tool 의 lock-in**: 1 ChatGPT 만 = λ§€ task 의 best fit X. - **Prompt κ°€ 1 try**: iterate κ°€ 핡심. - **AI 만 + critical thinking μ•ˆ 함**: λ§€ user 의 atrophy. - **Learning κ°€ 정적**: λ§€ 6 month 의 update ν•„μš”. - **AI ban (학ꡐ)**: 학생 의 reality 의 뢀적응. ## πŸ§ͺ 검증 μƒνƒœ (Validation) - **정보 μƒνƒœ:** verified (concept-level). - **좜처 신뒰도:** B (Long et al. 2020 paper, Ethan Mollick "Co-Intelligence", AI4ALL curriculum, UNESCO AI literacy framework). - **κ²€ν†  이유:** Manual cleanup. λ§€ framework κ°€ active. λ§€ 6 month review. ## 🧬 쀑볡 검사 (Duplicate Check) - **κΈ°μ‘΄ μœ μ‚¬ λ¬Έμ„œ:** [[Prompt_Engineering|Prompt-Engineering]] (subset), [[AI-Ethics]] (related), [[Critical Thinking]] (parent). - **처리 방식:** KEEP (overall framework). - **처리 이유:** Literacy κ°€ holistic. λ§€ component 의 own document. ## πŸ•“ λ³€κ²½ 이λ ₯ (Changelog) | λ‚ μ§œ | λ³€κ²½ λ‚΄μš© | 처리 방식 | 신뒰도 | |------|-----------|-----------|--------| | 2026-05-08 | P-Reinforce Phase 1 μ •κ·œν™” | UPDATE | A | | 2026-05-09 | Manual cleanup β€” pattern + curriculum + μ•ˆν‹°νŒ¨ν„΄ + λ§€ audience 의 priority | UPDATE | B |