AI and Recognition Technologies to Facilitate English as Foreign Language Writing for Supporting Personalization and Contextualization in Authentic Contexts

Abstract

English as Foreign Language (EFL) writing is challenging for students due to the lack of related lexical resources to inspire them for meaningful writing besides grammar feedback. Moreover, it is crucial to design personalized feedback based on students’ original writings in helping to improve their writing abilities. Therefore, we developed Smart RoamLingo app to help writing meaningful content in authentic contexts with sample sentences based on texts generated by recognition technologies named as AI-Sample Sentences (AI-SS) and writing qualities with revision suggestions based on personal original writings named as AI-Writing Feedback (AI-WF). 104 undergraduate students were assigned into an experimental group (EG) and two control groups. EG significantly outperformed the other groups in the post-test. In EG, AI-SS can help the contextualization to produce comprehensive content actively while AI-WF can help the personalization to improve the writing quality through several revisions. Hence, the cohesion and consistency of their writing can be enhanced. Further, the total number of revisions in practices and the assignment score can significantly predict the post-test. Moreover, EG students revealed that Smart RoamLingo was useful for EFL writing. Therefore, it is strongly suggested to use Smart RoamLingo with AI-SS and AI-WF for EFL writing in authentic contexts.