人工智慧在教育評量中的角色、影響與教師職責之探討

The Role, Influences, and Responsibilities of Teachers in Artificial Intelligence for Educational Assessment

本文探討人工智慧（AI）與生成式人工智慧（GAI）在教育評量中的應用現況、影響與教師職責轉變。AI擅長即時分析大規模學習數據，GAI則能快速生成試題與個別化回饋，兩者整合形成「生成與評改一體化」的新型評量模式，有效提升形成性評量的即時性與靈活性。然而，AI系統在透明性、數據偏誤、隱私保護與質性能力評估方面仍存限制，對教育公平與信任機制帶來影響。在此背景下，教師角色正從評分者轉型為AI監督者與學習設計者，需具備資料素養、評量設計素養與AI監督素養，並可參照TPACK與DigCompEdu框架實施三階段監督評量流程。本文建議未來研究應強化跨域整合、提升演算法可解釋性、完善隱私與倫理治理機制，並透過教師專業培訓與人機協作，建構具備效率、公平與人文關懷的智慧教育評量體系，實現永續教育的發展目標。

This study explores the current applications, influences, and evolving teacher responsibilities related to Artificial Intelligence (AI) and Generative Artificial Intelligence (GAI) in educational assessment. While AI excels in real-time analysis of large-scale learning data, GAI enables rapid generation of test items and personalized feedback. The integration of both technologies has given rise to a new assessment model that combines generation and evaluation, enhancing the immediacy and flexibility of formative assessment. However, issues such as algorithmic opacity, data bias, privacy concerns, and limitations in assessing qualitative competencies present significant influences to fairness and trust in education. Against this backdrop, the teacher’s role is shifting from that of a grader to an AI supervisor and learning designer, requiring competencies in data literacy, assessment design, and AI supervision. The study suggests leveraging the TPACK and DigCompEdu frameworks to implement a three-stage supervised assessment process. It further recommends that future research strengthen interdisciplinary integration, improve algorithm explainability, establish robust privacy and ethical governance mechanisms, and enhance teacher training and human–AI collaboration. Ultimately, the goal is to build an intelligent assessment ecosystem that is efficient, fair, and human-centered, aligned with the vision of sustainable education.