Factors influencing pre-service integrated science teachers’ acceptance and use of generative ai tools for learning at University of Education, Winneba

Abstract

Generative Artificial Intelligence (Gen AI) tools such as ChatGPT and Google Bard are increasingly transforming higher education worldwide. Their ability to support personalised learning, generate quick information, and enhance students’ academic productivity by simplifying complex tasks has drawn attention to their adoption within teacher education. Despite these opportunities, limited knowledge exists on the preservice integrated science teachers in the use of generative AI tools in Ghana. This study therefore examined factors influencing pre-service integrated science teachers’ acceptance and use of generative AI tools in learning at the University of Education, Winneba. Adopting a descriptive survey design, the study collected quantitative data from 300 pre-service integrated science teachers across levels 100 to 400. Descriptive statistics (frequencies, percentages, means, and standard deviations), reliability tests (McDonald’s Omega), and inferential statistics were used to analyse the quantitative data. The findings showed that 74% of the pre-service integrated science teachers demonstrated a strong behavioural intention to adopt generative AI tools, with an overall mean score of (Mean=3.67±1.03) on intention. However, only 49.3% reported frequent use. The results further indicated that performance expectancy (Mean=3.90±1.13), effort expectancy (Mean=3.70±0.98), hedonic motivation (Mean=3.73±1.02), and social influence (Mean=3.36±0.92) significantly predicted 74% variation in behavioural intention. Ethical concerns were moderately expressed with an overall mean (Mean=3.01±0.81), however, pre-service integrated science teachers worried about whether using generative AI to complete their work was morally acceptable (Mean=3.21±1.12), plagiarism (Mean=3.00±1.06), misinformation (Mean=3.18±1.11), and privacy breach (Mean=3.10±1.14). Pre-service integrated science teachers’ mitigation strategies included integrating digital ethics into teacher training programme (Mean=3.84±1.21) and creating and sharing clear policies on how to ethically use AI tools (Mean= 3.80±1.18) among others. The study concludes that while generative AI tools hold strong potential to enhance integrated science learning, their adoption is constrained by ethical dilemmas and absence of institutional guidelines. It is recommended that the University of Education, Winneba, and the Faculty of Science Education should develop clear policies, integrate ethical AI into teacher training curricula, and strengthen digital infrastructure to promote responsible and effective use.