Pre-service Teachers’ Discussion on ChatGPT: Using the LIWC

As the development of Artificial intelligence (AI) technologies has rapidly increased in education (Prahani et al., 2022), teachers are expected to adopt AI tools for effective teaching and learning. Acknowledging that early exposure to AI helps pre-service teachers apply AI to their teaching practice, we examined pre-service teachers’ perceptions and experiences of AI through linguistic features.

The study involved 54 pre-service teachers from a Southeastern university in the US. They engaged in asynchronous online courses and participated in weekly discussions. In this proposal, a class of eleven students’ discussion posts and comments were analyzed, and the full data will be presented in the paper. We uploaded a total of 428 sentences into the Linguistic Inquiry and Word Count (LIWC) software for linguistic analysis.

When asked about expectations and perceptions of AI in the first week, students used slightly more negative emotional expressions than positive ones. In the following week, they were requested to report any issues, challenges, or difficulties related to the AI-generated content. They seemed to assess the responses from ChatGPT positively based on the higher percentages of positive emotional expressions. Additionally, they used authentic language the most this week. In the third week, students explored effective strategies to use ChatGPT for task completion or problem-solving. As they approached this topic critically and logically, the use of emotional expressions was lower than in the other three weeks, while analytic language use was the highest. In the last week, they were asked to report any issues or concerns about AI ethics. They showed slightly more negative emotional language use than positive and linguistic language use the most. Overall, students used more analytical language than emotional expressions in their online discussion posts across the week. This study offers insights into the application of LIWC in online discourse analysis.