Artificial Intelligence in Biomedical Research: From AI Tools to AI Mega Agents in a Human-in-Charge Framework

Artificial intelligence (AI) is increasingly integrated across biomedical research workflows, from literature screening and image analysis to manuscript drafting and reference management; however, its current use remains largely fragmented into single-task tools. This narrative review summarizes the present role of AI in biomedical research and explores the future potential of AI mega-agents-integrated systems that coordinate specialized subagents under explicit human supervision. Evidence was obtained through a narrative search of PubMed, Scopus, Web of Science, Google Scholar, and Semantic Scholar using terms related to AI, biomedical research, large language models, AI agents, medical publishing, and governance, with priority given to peer-reviewed literature and ethical guidance from the International Committee of Medical Journal Editors, Committee on Publication Ethics, World Association of Medical Editors, World Health Organization, United Nations Educational, Scientific and Cultural Organization, the European Union Artificial Intelligence Act, and Indian Council of Medical Research guidelines. AI improves speed, scale, and reproducibility in literature review, protocol development, data analysis, writing, and figure generation, but also introduces risks, including hallucination, fabricated citations, bias, opacity, and inappropriate authorship claims. Emerging agentic AI systems may support project-level coordination of literature, protocols, statistics, writing, and ethics while preserving human accountability. We propose a human-in-charge AI research framework, operationalized through a biomedical research AI governance pyramid comprising technical robustness, workflow controls, institutional governance, and societal oversight. AI should remain a powerful copilot, not the captain of biomedical research.