批判媒體素養觀點下成人學習者運用生成式AI製作廣播節目的創作歷程分析

An Analysis of Adult Learners’Creative Process in Producing Radio Programs Using Generative AI From the Perspective of Critical Media Literacy

本研究以批判媒體素養為理論基礎，探討成人學習者在生成式AI輔助下參與廣播節目企劃與製作之創作歷程。採質性多重個案研究，以成人教育課程為場域，並結合影音製作流程與理論導向之分析面向進行分析。研究發現：第一，生成式AI由工具轉為創意中介與知識夥伴，支持主題發想與內容生成；第二，生成式AI促進批判媒體素養實踐，學員透過檢核、修潤與再詮釋，從技術使用者轉化為意義建構者；第三，生成式AI強化成人學習特徵，展現自主學習、情境調整、資源整合與反思遷移。理論上，本研究建構生成式AI輔助批判媒體素養學習模式；實務上，提出生成式AI融入廣播教學之四階段流程，並建議未來擴大樣本與進行量化驗證。

This study, grounded in the perspective of critical media literacy, explores the creative processes of adult learners engaging in radio program planning and production with the assistance of Generative Artificial Intelligence (Gen AI). A qualitative multiple-case study approach was adopted, situated within an adult education course. Data were collected from three participants, including their project proposals, scripts, final productions, and reflective reports. The analytical framework integrates four stages of media production—pre-production, recording, post-production, and distribution—along with two theory-driven matrices: the four core competencies of critical media literacy and key characteristics of adult learning. The findings reveal three main insights. First, Gen AI has evolved from a technical tool into a creative mediator and knowledge partner, supporting idea generation and content development. Second, Gen AI facilitates the practical enactment of critical media literacy, as learners critically evaluate, revise, and reinterpret AI-generated content, undergoing a transformation from technical users into meaning makers. Third, Gen AI enhances adult learning characteristics; including self-directed learning, situational adaptation, resource integration, and reflective transfer. Theoretically, this study proposes an AI-assisted critical media literacy learning model. Practically, it offers a four-stage instructional framework for integrating Gen AI into radio education and suggests future research to expand sample sizes and incorporate quantitative validation.