Practice, progress, and proficiency in sustainability, Год журнала: 2024, Номер unknown, С. 325 - 356
Опубликована: Июнь 28, 2024
Generative AI-driven automated essay scoring (AES) is expected to revolutionize personalized education by offering customized feedback students. However, the reliability of these systems currently undermined inherent limitations, such as tendency for “hallucination,” where AI generates factually incorrect or irrelevant information. To mitigate issues and bolster trustworthiness AES, this chapter argues that implementation explainable (XAI) crucial. Suitable XAI algorithms could make GenAI's decision-making process transparent, allowing educators students understand trust provided, thus ensuring effective integration in education. Furthermore, outlines several recommendations achieving a responsible GenAI-driven AES system.
Язык: Английский