Generative artificial intelligence and large language models in pharmaceutical formulation and personalized pharmacy: a review of opportunities, technical constraints, and regulatory readiness.

OBJECTIVE: This article reviews the emerging applications of generative artificial intelligence (GenAI) and large language models (LLMs) in areas beyond early drug discovery, specifically focusing on drug formulation and personalized pharmacy. SIGNIFICANCE OF REVIEW: This convergence supports more data-driven and potentially patient-specific formulation strategies. By combining the design capabilities of LLMs with the analytical power of generative models, researchers can address complex formulation challenges with increased precision. KEY FINDINGS: We analyze the impact of generative architectures (e.g. transformers, variational autoencoders) on formulation science, highlighting their ability to forecast critical properties such as solubility, stability, and excipient interactions. Furthermore, we investigate their role in the de novo design of nanocarriers and 3D-printed dosage forms. In the clinical setting, we examine how these models interpret genetic data to support the prediction of individual drug responses. CONCLUSION: The integration of GenAI and LLMs within pharmacy has the potential to accelerate development processes and pave the way for a new and practical approach to customized therapies. These developments may contribute to the future integration of personalized formulation strategies in pharmaceutical development, contingent upon robust validation and regulatory alignment.