The rapid evolution of generative Artificial Intelligence (AI) tools such as ChatGPT presents new opportunities for enhancing productivity, creativity, and decision making within public sector institutions. However, limited empirical research exists o...
The rapid evolution of generative Artificial Intelligence (AI) tools such as ChatGPT presents new opportunities for enhancing productivity, creativity, and decision making within public sector institutions. However, limited empirical research exists on how public employees in developing country contexts perceive and adopt these emerging technologies. This study investigated the determinants of public administrators’ behavioral intention to adopt ChatGPT with Kenya’s Ministry of Energy and Petroleum (MoEP), employing a mixed-methods sequential explanatory design. The quantitative phase used the Unified Theory of Acceptance and Use of Technology (UTAUT) to examine the influence of Performance Expectancy (PE), Effort Expectancy (EE), Social Influence (SI), Facilitating Conditions (FC), and Personal Innovativeness (PI) on Behavioral Intention (BI). Survey data (n=187) were analyzed using multiple regression techniques.
The results revealed that PE and PI had significant positives effects, indicating that employees are more likely to adopt ChatGPT when they perceive it as useful and are personally inclined toward exploring new technologies. Conversely, EE had a significant negative effect, suggesting that concerns regarding accuracy, data privacy, complexity, and limited contextual relevance reduce employees’ willingness to use ChatGPT. SI and FC were found to be statistically insignificant, reflecting limited leadership direction and insufficient institutional infrastructure to support AI adoption.
The qualitative phase, conducted through follow up interviews and thematic analysis (Braun & Clarke, 2006), elaborated four dominant themes: (1) utilization of ChatGPT in daily work settings, (2) creativity and innovation enhancement, (3) adoption-related challenges and barriers, and (4) leadership and institutional readiness. Participants described ChatGPT as a valuable tool for drafting reports, preparing briefs, simplifying technical concepts, and generating ideas, reinforcing the significant influence of PE. They also highlighted the role of personal curiosity and experimentation in shaping adoption behaviors, consistent with the strong significance of PI. At the same time, concerns over data privacy, inaccuracies, and account limitations aligned with the negative EE effect. The absence of clear organizational policies, limited AI skills, and lack of secure platforms explained the insignificance of SI and FC in the quantitative results.
Overall, the study concludes that whereas AI and specifically ChatGPT in Kenya’s public service sector context is still in nascent stage and that usage is still at a superficial level rather than complex and advanced, generative AI holds substantial potential to improve administrative efficiency, analytical capacity and creativity within Kenya’s public sector. However, meaningful and responsible adoption will require structured leadership support, sector-specific training, secure AI infrastructure and clear policy frameworks that safeguard data and guide ethical use. The findings provide evidence-based insights that can inform the implementation of Kenya’s Digital Master Plan (2022-2032) and the National AI Strategy (2025-2030), contributing to ongoing efforts towards AI-Augmented, digitally transformed and globally competitive public service.