生成式AI輔助快速設計之後設認知歷程探討

Metacognitive Processes in Generative AI-Supported Rapid Design Tasks

隨著生成式人工智慧在設計領域的應用日益普及，設計者與工具之間的互動關係隨之轉變，尤其在思考歷程與認知策略上的表現愈趨複雜。本研究旨在探究生成式AI（generative artificial intelligence, GAI）輔助快速設計歷程中，設計者所展現之後設認知行為與策略調整。研究採質性個案研究法，邀請4位綜合型大學工業設計研究所碩士生執行1至1.5小時的快速設計任務。任務結束後進行回溯式放聲思考訪談，並針對逐字稿進行編碼與分析。編碼架構分為兩大面向：「設計階段」包含探索（discover）、定義（define）與設計（design）；「設計的後設認知」則涵蓋後設認知知識、後設認知監控與後設認知調節三類。研究結果指出，生成式AI主要應用於探索階段（Dc）與設計階段（Ds），輔助「設計前期」與「發散導向」相關任務。在後設認知運作方面，AI工具提供的資訊與生成內容促進設計者反覆測試與快速調整，形成「高頻迭代」的設計模式，對於提升成果品質與縮短設計時程皆具潛在效益，扮演「驅動設計行為」與「創造反思機會」的角色。

As generative artificial intelligence (GAI) becomes increasingly integrated into design practices, the dynamics between designers and tools are evolving—particularly in ways that increase the complexity of cognitive strategies and reflective processes. This study aims to investigate the metacognitive behaviors and strategic adjustments exhibited by designers during generative AI-supported rapid design processes. Adopting a qualitative case study approach, the study recruited four graduate students from the Department of Industrial Design at a comprehensive university in Taiwan. Each participant used generative AI tools to complete a rapid design task lasting approximately 1-1.5 hours. Following the task, retrospective think-aloud interviews were conducted to elicit participants’design reasoning and reflective thought. Coding and analysis were performed on verbatim transcripts using a scheme comprising two dimensions: (1) Design Stages—discover (Dc), define, and design (Ds); and (2) Metacognition of Design—metacognitive knowledge, metacognitive monitoring, and metacognitive regulation. Findings revealed that generative AI was primarily applied during the discover and design stages, supporting pre-design and divergence-oriented tasks. In addition, the AI tools facilitated frequent cycles of reflection and adaptation, forming what can be described as a“high-frequency iteration”pattern. This pattern offers potential benefits for enhancing output quality and shortening design cycles, with AI playing the roles of driving design behavior and creating opportunities for reflection. The study highlights AI’s dual function as a catalyst for design action and a source of reflective opportunities.