AI and Machine Learning in Public Health: Ensuring Algorithmic Fairness and Ethical Data Utilization for Population Health Management

AI and ML are revolutionizing public health by advancing predictive analytics, optimizing health resource deployment, and ensuring ethical, fair decision making and many other use cases. This dissertation investigates the use of AI for algorithmic fairness and ethical use of data for population health management. Four of the main ML algorithms, Random Forest, Support Vector Machine (SVM), Neural Networks, and Fair Federated Learning (FFL), have also been implemented and compared in terms of predictive accuracy of disease diagnosis and risk quantification. Results from the experimental results of Neural Networks showed the highest accuracy (94.2%), while SVM produced an accuracy of 91.6%, Random Forest was 89.4% accuracy, and the final was FFL of 87.3% accuracy. And all of these showed the effectiveness of deep learning models. Fairness metrics also showed FFL helps reduce bias by 22% compared to traditional centralized models and produces fair healthcare outcomes across different populations. The necessity of ethical AI protocols, privacy of data, fairness conscious algorithms in public health use cases is again emphasized in the research. To make AI deployment in healthcare a responsible practice, more work in the direction of making models more interpretable, methods to mitigate bias, compliance to regulation is required. The power of these solutions with AI in developing more open, fair and effective public health administration systems is highlighted in these results.