An Empirical Research on Using Generative Artificial Intelligence to Assist Primary School Students in Hong Kong Learning Chinese Expository Writing

Abstract

<p>Research shows that many students dislike writing and often encounter various difficulties, such as writer's block, limited vocabulary, grammatical errors, poor sentence structures, difficulties in organising ideas, lack of clarity, and inability to engage readers effectively. On the other hand, teachers face hurdles in assisting students to apply their reading skills to their writing and addressing the unique writing needs of each student. Due to heavy workloads, teachers may require up to three weeks to correct students' compositions, hindering the provision of timely feedback and discouraging frequent writing practice for improvement.<br></p><p>The research team developed an online Chinese writing assessment platform named the “mAI Mind” by applying the instant content analytic and idea generation functions of “Generative Artificial Intelligence” and referencing the genre theory, writing curriculum framework and assessment criteria of Hong Kong primary and secondary schools. 176 Grade 6 students were recruited to have a trial run of the system, to examine its effectiveness in addressing the challenges in writing education and providing personalised writing suggestions for students by adapting quasi-experimental research design. The results of the student questionnaire show that mAI Mind significantly improves students' writing attitudes and motivation. Content analysis of their pre- and post-writing observed diversified vocabulary usage in a short period of time, leading to long-term enhancement of their writing abilities. The system’s data analysis capabilities not only alleviate teachers' assessment pressures but also enhance the focus and depth of writing feedback. It opens a new path for the learning and teaching of writing.</p>

研究顯示，很多學生不喜歡寫作，寫作時經常遇到各種困難，例如不知道寫甚麼、詞彙量不足、語法錯誤、句意不清、內容欠缺組織，以及缺乏讀者意識等。老師幫助學生應用從讀文教學中獲得的技能到寫作時，往往需要付出很大的努力，並且難以滿足每位學生的寫作需要。此外，由於日常工作量大，老師可能需要三星期才能完成批改學生的作文，這使他們無法及時給予學生寫作回饋，也難以讓學生通過多練習來提升寫作水準。

研究團隊應用「生成式人工智慧」能夠快速分析文字和生成意念的特點，根據文類理論、香港中小學的寫作教學目標和評改標準，自主研發名為「mAI Mind」的線上中文寫作評改平台。團隊邀請了一所小學共176位小六學生利用該系統進行准實驗研究，檢視其對解決實用文寫作學與教的難點，為學生提供個人化的寫作建議。問卷調查結果顯示該系統能於短時間內有效改善學生的寫作態度和動機，文本分析學生的前後測寫作表現，發現他們的詞語運用有明顯進步。長遠而言，對提升學生寫作能力有正面影響。該系統的資料分析功能不但能減輕老師的評改壓力，而且能提高寫作評改的聚焦性和深度，為寫作學與教開創新路徑。