Speakers
Description
This research-oriented presentation will introduce the new Auto-Peer’s customizations for Japanese foundational level learners, providing a data-informed compliment to our practically-oriented workshop. We will describe the desirable aspects of generative AI (GenAI) output that we leveraged to help students further improve their academic writing. Specifically, we describe how analysis of the discursive features of Japanese (n=68) and UAE (n=81) student texts when compared to GenAI-generated texts demonstrated insufficient cohesion and, for the Japanese students, a lack of consistent verb use. We explain how, using these insights, the new Auto-Peer leverages these more proficient aspects of GenAI writing to ethically inform students about how they could improve their writing, leaving them in charge of and responsible for their texts.