ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING FOR DIABETES PREDICTION AND MANAGEMENT: STRENGTHENING AI-DRIVEN HEALTHCARE INNOVATION AND BIG DATA SECURITY WITHIN THE U.S. PUBLIC HEALTH INFRASTRUCTURE

Abstract

This study addresses the challenge of ensuring that artificial intelligence and machine learning for diabetes prediction and management are deployed within secure big data environments in the U.S. public health infrastructure. The purpose is to quantify how AI and machine learning adoption, big data security practices, healthcare innovation, and trust in AI systems jointly predict diabetes management effectiveness in cloud and enterprise settings. A quantitative cross sectional, case-based survey design was applied to 285 professionals from 18 U.S. healthcare organizations operating EHR and analytics platforms on enterprise or cloud infrastructures. Key latent variables were AI or machine learning adoption for diabetes, big data security, healthcare innovation, trust in AI, and diabetes management effectiveness, all measured on five-point Likert scales and analyzed using descriptive statistics, correlations, and multiple regression with moderation. AI and machine learning adoption showed a strong positive association with diabetes management effectiveness (r = .52, p < .001; β = .29, p < .001) and healthcare innovation (r = .63, p < .001; β = .63, p < .001). Big data security strongly predicted trust in AI (r = .57, p < .001; β = .57, p < .001) and significantly moderated the adoption effect on effectiveness (interaction β = .14, ΔR² = .02, p = .005). The full regression model explained 60 percent of the variance in diabetes management effectiveness (R² = .60), demonstrating that secure and trusted AI enabled environments are critical for translating predictive and decision support tools into tangible improvements in population-oriented diabetes care.