Advancements in translation software and the application of machine learning through neural networks have sparked a revolution in language education. However, they also pose challenges to traditional homework and assessment methods. The advent of machine translation provides ESL students with the ability to generate English texts without relying on English language skills. This emerging reality challenges the credibility of English writing assessments. This paper suggests an innovative approach to writing tasks and assessment, aiming to ensure that the evaluated work genuinely reflects students’ own efforts while allowing them to incorporate technology in a way that enriches rather than hinders learning. The proposed method has two phases. Students write a first draft without access to technology. In the subsequent technology-assisted phase, students transcribe their handwritten work as a first draft and create a second draft using machine translation to address errors in grammar, spelling and word-choice. Finally, students are tasked with highlighting the disparities between the two drafts.
Paul McKenna (Tue,) studied this question.