Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{192062,
        author = {Sourabh Bagoria and Rohit Namdeo Ingole and Bommareddy Devi Aruna Jyothi and Aamir Patel and Daizy Tak and Priyanka Rajendra Wadile and Cheruku Akshith and Sumayya Sanober},
        title = {The Role of Artificial Intelligence in Pharmacovigilance: A Narrative Review of Methods, Performance and Limitations},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {8},
        pages = {8842-8859},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=192062},
        abstract = {Pharmacovigilance represents a cornerstone of pharmaceutical regulation and clinical governance, serving as a critical mechanism for the continuous evaluation of the safety profile of medicinal products following their authorization for public use. While pre-marketing clinical trials provide essential evidence regarding efficacy and short-term safety, their methodological constraints—such as limited sample sizes, controlled study environments, restricted patient heterogeneity, and relatively short durations of exposure—inevitably hinder the identification of rare, delayed, cumulative, or context-dependent adverse drug reactions that may only emerge under real-world conditions. Consequently, post-marketing pharmacovigilance systems are tasked with the complex responsibility of monitoring drug safety across diverse populations, therapeutic indications, and healthcare settings over extended periods of time. However, traditional pharmacovigilance frameworks, which rely predominantly on spontaneous reporting systems, manual case processing, and rule-based statistical signal detection methodologies, are increasingly challenged by the exponential growth in the volume, velocity, and heterogeneity of safety-related data generated in modern healthcare ecosystems. These limitations have resulted in delayed signal detection, substantial operational burden, variability in case quality assessment, and reduced capacity to extract meaningful insights from unstructured or non-traditional data sources such as electronic health records, biomedical literature, and patient-generated digital content.
In this context, artificial intelligence has emerged as a potentially transformative paradigm capable of addressing many of the structural inefficiencies inherent in conventional pharmacovigilance practices. Artificial intelligence-driven methodologies, encompassing machine learning, deep learning, and natural language processing techniques, enable automated ingestion and analysis of large-scale, heterogeneous datasets, facilitate advanced pattern recognition, and support adaptive learning from continuously evolving data streams. Within pharmacovigilance operations, these technologies have been applied to a wide range of functions, including automated adverse event case intake, medical coding and normalization using standardized terminologies such as MedDRA, signal detection and prioritization, literature surveillance, and benefit–risk assessment. Early evidence suggests that artificial intelligence-based systems can enhance processing efficiency, improve consistency in case assessment, and increase sensitivity for early safety signal identification when compared with traditional disproportionality analyses. Nevertheless, the deployment of artificial intelligence in pharmacovigilance is accompanied by significant methodological, regulatory, and ethical challenges, including data quality and representativeness limitations, algorithmic bias, lack of transparency and interpretability in complex models, constrained generalizability across populations and therapeutic areas, and the continued necessity for expert human oversight to ensure clinical relevance and regulatory compliance.
This narrative review critically examines the role of artificial intelligence in pharmacovigilance by synthesizing current methodological approaches, evaluating reported performance outcomes, and identifying key limitations that influence real-world implementation. Furthermore, it explores emerging regulatory perspectives and future directions aimed at fostering responsible integration of artificial intelligence into pharmacovigilance systems. By providing a comprehensive and clinically grounded assessment, this review seeks to contribute to the evolving discourse on how artificial intelligence can be strategically leveraged to augment, rather than replace, traditional pharmacovigilance practices, ultimately supporting more timely, robust, and patient-centered drug safety surveillance.},
        keywords = {Artificial Intelligence; Pharmacovigilance; Adverse Drug Reactions; Drug Safety Surveillance; Machine Learning; Deep Learning; Natural Language Processing; Signal Detection; Real-World Evidence; Post-Marketing Surveillance; Regulatory Science; Benefit–Risk Assessment},
        month = {January},
        }