Speaker
Description
This research focuses on the perceptions of the English majors at HUFLIT about the efficacy of ChatGPT in improving their reading comprehension. With the use of the convergent parallel mixed methods design, a total of 207 students were surveyed. In order to balance both depth and breadth, a total of 144 students participated in a quantitative survey using Likert scales (analyzed using descriptive statistics), while another group of 63 students provided qualitative data through the open-ended questionnaire method (analyzed utilizing thematic analysis). Triangulated findings show that the students perceive the utility of ChatGPT as a “Linguistic Equalizer,” making them think of it as highly useful in decoding the complexity of syntax (Mean=3.96) and symbolism (Mean=3.90) in a bid to overcome academic density. However, a dependency paradox was observed with regards to how students feel about ChatGPT where despite having a high perception of personal growth in reading skills (Mean=3.97), their independent reading skills have worsened due to reliance on the tool (Mean=2.44). Moreover, due to ChatGPT’s inability to exhibit emotional intelligence, an "emotional ceiling" was noted, contributing towards creating a form of skeptical trust due to the cross-check of its output. It is concluded that even though ChatGPT works as a cognitive scaffold for students, it must not replace independent reading. The study suggests that ChatGPT can be seen as a good cognitive scaffold, but universities should conduct training on AI literacy.
Keywords: ChatGPT, reading comprehension, English majors, mixed methods
