Artificial Intelligence and Science Technology Innovation Mainstreaming in Kenyan Public Research Institutions

Abstract

This study examined the influence of Artificial Intelligence (AI) adoption on the mainstreaming of Science, Technology, and Innovation (STI) in Kenyan public research institutions. It assessed four AI dimensions; Algorithmic Models, AI Software Stacks, AI Infrastructure, and Expertise as independent variables, with STI mainstreaming as the dependent construct. Data was collected through a census of 11 public research institutions using structured questionnaires and interviews. Exploratory Factor Analysis (EFA) was used to identify underlying factor structures, while Confirmatory Factor Analysis (CFA) validated the measurement model. Structural Equation Modeling (SEM) then tested the hypothesized relationships. The results revealed significant positive associations between AI adoption and STI mainstreaming. Algorithmic Models emerged as the strongest predictor (β = 0.62, p = 0.001), followed by Expertise (β = 0.56, p = 0.002), AI Software Stacks (β = 0.47, p = 0.003), and AI Infrastructure (β = 0.41, p = 0.012). These findings underscore the critical role of AI in enhancing STI through improved decision-making, operational efficiency, and innovation capacity. However, challenges such as infrastructure gaps and limited expertise were identified as barriers. The study concludes that AI is a key enabler of STI mainstreaming, and its successful integration requires strategic investment in infrastructure, capacity building, and human capital. The findings provide a solid empirical foundation for future research on AI's role in advancing STI and sustainable development.