Sessions /
Effects of Using AI and PeerEval in Teaching EFL #112

The focus of this study was to investigate the use of AI speakers and the effectiveness of PeerEval software in improving English proficiency. The longitudinal investigation was carried out from April 2019 to January 2020 using Alexa, PeerEval, ATR CALL Brix, Facebook, Line, and online materials. The participants were 59 first-year students who were engaged in flipped learning lessons through CLIL for two semesters. Pre-test and post-test TOEIC scores, as well as post-training survey results, were used in evaluating the overall effectiveness of two groups. The participants were divided into two groups, one with using AI speakers (n=29) and the other with non-AI speakers (n=30). Non-AI participants were engaged in e-learning outside of the class. While the AI group is studying English with the AI speakers, participants recorded short movie clips of their learning experiences. AI group participants also kept written diaries with their observations about the contents and duration of their studies, periodically recording their thoughts using a smartphone. AI group indicated the improvement of TOEIC mean score from 407 (SD:113) to 604 (SD:92), and the non-AI group showed the increase from 447 (SD:93) to 598 (SD:147). Those who studied English with AI speakers outside of the classroom outperformed the group with non-users of AI speakers on a post-course TOEIC test. This research included some limitations in attempting to ascertain the effectiveness of using AI in teaching language and communication by exclusively focusing on the role of AI and students’ English skills.