2022 Fiscal Year Research-status Report
Investigating an AI-based writing assistant's impact on English language learners' writing proficiency
Project/Area Number |
22K00718
|
Research Institution | University of Hyogo |
Principal Investigator |
ガイエド ジョンモリース 兵庫県立大学, 国際交流機構, 英語インストラクター (20868393)
|
Co-Investigator(Kenkyū-buntansha) |
クロス ジェフリーS 東京工業大学, 環境・社会理工学院, 教授 (90532044)
|
Project Period (FY) |
2022-04-01 – 2025-03-31
|
Keywords | CALL / AI assistant / L2 writing / metacognition / artificial intelligence / AIED |
Outline of Annual Research Achievements |
Progress thus far has been encouraging on the development and testing of an AI based writing assistant for second language learners. Three peer reviewed papers were published in international journals, in addition to disseminating the results at international conferences. These results demonstrate the effectiveness and innovation of the approach used in this research. Progress in developing the core features and functionalities of the writing assistant were also expanded to include better functionality and meta-cognitive support to the L2 writer. In addition, further collaboration opportunities after presented themselves due to the publication results of this research. The researcher is very encouraged with the recent interest in AI generated content such as ChatGPT, Google Bard and etc. The research results show promise and further development of AI-based assistance for L2 leaners is warranted.
|
Current Status of Research Progress |
Current Status of Research Progress
1: Research has progressed more than it was originally planned.
Reason
The researcher is still continuing to disseminate the results of this research. This includes further participation in international second language research conferences. Additionally, the researcher is creating a framework for the use of AI in the classroom. The intention is to circulate a framework that educators from different fields can adapt and use in their classrooms.
|
Strategy for Future Research Activity |
Due to the rapid development of large language models (LLMs) such as ChatGPT and Google Bard, the researcher is reassessing the approach and base technology used in this research. This research uses technology largely based on GPT-2 language models which have been made obsolete in the span of a few short years. The language support chat-based models (such as ChatGPT which uses GPT-4) show promise of providing more sophisticated and understandable feedback to the second language writer. Therefore, the researcher is investigating updating the underlying technology used to improve its overall performance.
|
Causes of Carryover |
Due to the COVID-19 pandemic, the number of conferences attended was very small. In addition development work on the software will be accelerated in the 2023 fiscal year in addition to recruiting more participants for data analysis.
|