Agentic GenAI for Infectious Disease Management: A Comprehensive Review

The global management of infectious diseases, from pandemics to antimicrobial resistance, remains a critical public health challenge. This comprehensive review paper synthesizes the emerging paradigm of Agentic Artificial Intelligence (AI) for infectious disease management, marking a significant evolution beyond traditional generative AI. We define Agentic AI as autonomous systems capable of reasoning, planning, and executing complex, multi-step tasks by leveraging tools such as scientific databases and analytical engines. The core architectural components—planning modules, tool use APIs, memory, and guardrails—are detailed, alongside examples from industry platforms like Oracle OCI and IQVIA. A systematic analysis demonstrates key applications: revolutionizing disease surveillance and forecasting with superior predictive accuracy; drastically accelerating antibiotic discovery through \textit{de novo} molecular design; augmenting clinical diagnostics and decision support; and automating scientific literature synthesis. The review further categorizes agent-specific approaches tailored to pathogen characteristics, including RNA viruses, drug-resistant bacteria, and neglected diseases. However, this promise is tempered by substantial challenges, including data bias, model hallucination, security vulnerabilities, and a lack of regulatory frameworks. Performance must be evaluated through multifaceted metrics like Task Success Rate and Medical Harmfulness Score, not just accuracy. A systematic exploration of key applications is presented, including enhanced disease surveillance and forecasting, accelerated drug and antibiotic discovery, AI-augmented clinical diagnostics and decision support, and automated scientific research. We further analyze the significant technical, ethical, and implementation challenges, such as data quality, hallucination risks, and the ``black box'' problem. Finally, we outline future directions, emphasizing the need for robust validation frameworks, human-AI collaboration models, and sustainable integration into public health infrastructure. The future direction emphasizes human-AI collaboration, robust benchmarking, and equitable deployment to avoid exacerbating global health disparities.