As generative AI tools such as ChatGPT are increasingly integrated into education, it has become crucial to understand their impact on students. This study examines how college students' self-efficacy in using ChatGPT influences their learning motivation and expectancy for success. Using a questionnaire, the study investigated students' confidence and ability in using ChatGPT, as well as whether it enhances their learning engagement. Participants were students from a university in northern Taiwan, and a total of 513 valid responses were collected. Females accounted for the majority (66.5%). By college affiliation, 53% of respondents were from the College of Humanities, Social Sciences, and Education. In terms of year level, first-year students comprised the largest group (29.2%), while seniors were the smallest group, accounting for only 16.6%.
This study analyzed the effect of ChatGPT self-efficacy on college students' learning motivation and expectancy for success, with self-efficacy as the independent variable and learning motivation and expectancy for success as the dependent variables. The results showed that students' self-efficacy positively influenced both learning motivation and expectancy for success. Moreover, male students scored significantly higher than female students in self-efficacy, learning motivation, and expectancy for success. Possible explanations include that males tend to demonstrate higher self-efficacy when facing new technologies such as ChatGPT, and may show greater motivation due to being more willing to experiment with new tools. In contrast, female students, even when performing well, may underestimate themselves (lower self-efficacy), whereas male students are more likely to overestimate or affirm their abilities, which in turn affects their expectancy for success.