A Digital-Ethical CMO model of research culture among STEAM faculty in Philippine SUCs

Abstract This paper examines how the research culture of Science, Technology, Engineering, Agriculture, and Mathematics (STEAM) faculty in Philippine SUCs is shaped by ability, motivation, and opportunity amid rapid AI-driven digital transformation. It extends the traditional CMO model into a Digital-Ethical CMO framework by integrating AI literacy and verification practices into faculty capability and embedding digital ethics and fair governance conditions into institutional opportunity. Data were drawn from 272 STEAM faculty across five SUCs using a convergent parallel mixed-methods design combining survey, interviews, and focus group discussions. Quantitative analyses used correlations, multiple regression, and bootstrapped mediation, and qualitative themes were developed through CMO-guided thematic analysis. Findings show that capability and opportunity have the strongest multivariable associations with research engagement, while capability has a modest direct association with research productivity. Opportunity shows a strong bivariate association with productivity, but multivariable and mediation tests do not support opportunity as a consistent pathway to productivity once capability is accounted for; productivity is more proximally associated with faculty capability, with opportunity operating primarily as an enabling participation context. Motivation relates to outcomes indirectly through collegial trust, recognition, and scholarly purpose. Qualitative themes highlight mentoring, digital equity, and ethical governance as conditions that strengthen participation and integrity in AI-enabled research. The study offers policy directions for Global South universities seeking inclusive and ethically grounded research cultures. These results position the Digital-Ethical CMO extension as a useful lens for aligning capacity-building, responsible AI governance, and equitable infrastructure with faculty engagement and output in resource-constrained settings.