Perceived Influence of Conversational AIs on Research Learning and Knowledge Acquisition among Science Education Undergraduates, Benue State University, Makurdi

This study investigated the perceived influence of conversational AIs on research learning and research knowledge acquisition among 400-level Science Education undergraduates at Benue State University, Makurdi using a descriptive comparative survey design. The study was anchored on Social Constructivism, Connectivism, the Technology Acceptance Model and the Unified Theory of Acceptance and Use of Technology. From a population of 124 400-level undergraduates in the Department of Science and Mathematics Education (2024/2025), a sample size of 95 (50 males and 45 females) was determined using Taro Yamane formula and drawn by a multi-stage sampling. Data were collected using Conversational AIs and Science Education Research Questionnaire (CAISERQ) and Research Knowledge Acquisition Test (RKAT). The instruments were validated by three experts and trial-tested with reliability values of 0.75 (Cronbach's alpha) and 0.78 (KR-20), respectively. Mean and standard deviation were used to answer research questions with a criterion mean of 2.50, while the independent samples t-test was used to test the hypotheses at 0.05 level of significance. Findings showed high perceived influence of conversational AIs on research learning for both genders, with all 15 items above the criterion mean and cluster means of 3.28 for males and 3.31 for females. The gender difference was not significant [t (93) = 0.756, p = .451]. Likewise, male (M = 23.06) and female (M = 23.02) undergraduates had similar research knowledge acquisition scores, with no significant difference [t (93) = 0.068, p = 0.946]. It was concluded that male and female undergraduates perceive and benefit from conversational AI in a similar way. The study recommended integration of AI literacy into research methodology courses and implementation of conversational AIs as supplementary supervision support.