運用 PDCA 導向學習模組提升與生成式 AI 協作程式設計成效：錯誤識別與修正策略能力培養;A PDCA-oriented instructional model for enhancing generative AI-assisted programming: developing error identification and repair strategy competence

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/98323

Please use this identifier to cite or link to this item: https://ir.lib.ncu.edu.tw/handle/987654321/98323

Title:	運用 PDCA 導向學習模組提升與生成式 AI 協作程式設計成效：錯誤識別與修正策略能力培養;A PDCA-oriented instructional model for enhancing generative AI-assisted programming: developing error identification and repair strategy competence
Authors:	蔡方慈;Tsai, Fang-tzu
Contributors:	資訊工程學系
Keywords:	生成式人工智慧協作程式設計;PDCA 循環;錯誤識別能力;修正策略能力;Collaborative programming with generative AI;PDCA cycle;Error identification;Repair strategy competence
Date:	2025-07-24
Issue Date:	2025-10-17 12:37:48 (UTC+8)
Publisher:	國立中央大學
Abstract:	隨著生成式人工智慧(Generative AI)技術的迅速發展，越來越多使用者透過生成式AI工具(如 ChatGPT)進行程式設計協作，從中獲得便利與效率。然而，對於非資訊背景或初學者而言，缺乏系統性學習策略與後設認知能力，可能導致在使用 AI 工具過程中出現錯誤依賴、無效驗證與錯誤理解等問題，進而影響程式設計的學習成效與問題解決能力。既有研究多聚焦於探討生成式AI所帶來的挑戰與偏誤，但對於如何透過教學設計系統性引導使用者與AI有效協作，仍屬發展初期。為此，本研究提出一套結合PDCA (Plan–Do–Check–Act)循環理論的教學模組，搭配情境式問題與案例式學習策略，協助學習者在與生成式AI協作程式設計過程中，培養其應具備的核心能力。本研究聚焦於兩項關鍵能力：錯誤識別能力與有效修正策略能力，並探討如何透過PDCA循環方式，強化學習者對於AI工具生成結果的理解、判斷與調整能力。期望透過此教學模組的導入，提升學習者在AI協作環境下的程式設計成效，最終目標在於培養初學者與專業開發者在問題解決能力上的落差，建立具備批判性錯誤識別能力與具體策略性修正能力的AI協作程式設計素養。 ;With the rapid advancement of generative artificial intelligence (Generative AI), an increasing number of users are engaging in collaborative programming through large language models (LLMs) such as ChatGPT. While these tools offer convenience and efficiency, novice programmers—especially those without a computer science background—often lack systematic learning strategies and metacognitive skills. This deficiency can lead to issues such as over-reliance on AI outputs, ineffective debugging, inadequate validation, and misinterpretation of results, ultimately hindering learning outcomes and problem-solving performance. Although prior studies have explored the challenges posed by generative AI, limited research has focused on instructional designs that guide learners to collaborate effectively with AI systems. To address this gap, this study proposes an instructional model grounded in the PDCA (Plan–Do–Check–Act) cycle, integrated with scenario-based problems and case-based learning strategies. The model aims to support learners in acquiring essential competencies for AI-assisted programming, with a particular focus on two core abilities: error identification and repair strategy development. Through iterative feedback and structured task engagement, the model fosters learners’ capacity to interpret, evaluate, and refine AI-generated code. This study seeks to enhance learners’ programming performance in collaborative programming with generative AI by implementing by a PDCA-oriented instructional model. Through structured guidance, the model seeks to bridge the gap between novices and expert developers in terms of problem-solving competence. The ultimate goal is to foster learners’ critical abilities in identifying errors and applying concrete, strategic repair approaches, thereby developing programming literacy essential for effective collaboration with generative AI.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	12	View/Open

社群 sharing

Loading...