ADAPTING ESL ASSESSMENT: STRATEGIES FOR MANAGING USE OF AI AND MACHINE TRANSLATION IN SPEAKING TASKS

Abstract

All educators must adapt to the development of technology. A recent development in language education is students’ use of artificial intelligence and machine translation to complete out-of-class assessment tasks. While this technology offers remarkable possibilities for language learning, it is very tempting for students to bypass the effort of creating assignments themselves and rely entirely on AI and machine translation. Also of concern is that students who use such technology may receive higher evaluations and better grades than those who don’t, as AI and translation software is now capable of producing language free from mistakes such as grammatical errors. This research focused on adapting both the instructions and grading rubrics of recorded speaking tasks, used for formative assessment in first- and second-year compulsory ESL classes at a university in Japan, to mitigate some of the concerns mentioned above. While it is not possible to completely eliminate the problem of AI-based cheating, this presentation will include some practical techniques that can reduce it to some extent.