Determinants of ChatGPT adoption in underdeveloped academic settings: a case study of Somalia using an extended TAM approach

Artificial Intelligence (AI) has revolutionized various domains, including higher education, particularly through generative tools like ChatGPT. To date, only one study has used the Technology Acceptance Model (TAM) to examine students’ intention to use ChatGPT in Somalia, focusing mainly on core constructs like perceived usefulness and ease of use. This study extends previous work by introducing additional constructs, including awareness of ChatGPT, facilitating conditions, and the moderating effects of trust and gender, uncovering more meaningful, localized determinants of ChatGPT adoption among university students in Mogadishu. A self-administered, cross-sectional online survey was conducted with 468 students from five universities in Mogadishu. Data were analyzed using SmartPLS 4. Results indicate that awareness alone does not directly impact adoption intention; however, perceived usefulness fully mediates this relationship. Further, social influence and facilitating conditions emerged as strong predictors of adoption intention, while perceived ease of use, trust, and gender differences showed weaker effects. The findings highlight the importance of peer influence, resource availability, institutional support, and perceived benefits in driving AI adoption in low-income settings. The study recommends that universities should clearly communicate ChatGPT’s practical benefits and promote peer-driven learning environments. The Somali higher education sector should establish and implement guidelines to ensure ethical and effective integration of AI tools like ChatGPT.