The integration of artificial intelligence (AI) and machine learning (ML) into pharmacovigilance (PV) marks a transformative step towards enhancing drug safety and patient outcomes. With their unparalleled ability to analyze vast and complex datasets, AI/ML technologies offer unprecedented opportunities to revolutionize adverse drug event (ADE) monitoring, signal detection, and causality assessment. This editorial explores the state-of-the-art applications, key challenges, and actionable strategies to unlock the full potential of AI in PV.

The applications of AI and ML methodologies have the potential to optimize drug safety monitoring by automating the analysis of adverse signals from various sources, such as health records, spontaneous and active surveillance techniques.

AI/ML has been shown to be useful for multiple aspects of data ingestion and for assessment of reported causality in individual case safety reports (ICSRs) [1]. Several ML techniques, such as Lasso shrinkage regression [2], Bayesian borrowing algorithms [3] and temporal scan statistics [4], have been shown to be applicable for signal detection. ML is useful for identification of syndromal patterns in data monitoring and performing complex modeling PV tasks [5].

Very recently, Zou et al. [6] demonstrated trends of utilizing the AI and ML derived from the published articles. These trends include highlight the expected wide applicability of the large and Natural Language models in the field of drug safety. They indicated the need of the concept of multiple data processing modalities and the automation of the different PV tasks. A hybrid model of AI and ML was suggested as the best solution to overcome the challenges facing this integration [7,8].

There are evidences of the applicability of AI in the life cycle of safe and effective drug development. Bhattamisra et al. [9] demonstrated the effectiveness of AI algorithms in enhancing patients’ safety outcomes in several case studies. Wanika et al. [10] utilized ML to aid in monitoring and management of cancer patients subjected to increased risk of serious adverse events. Chandak and Tatonetti [11] used ML technology to develop an “AwareDX” (Analyzing Women At Risk for Experiencing Drug toxicity) algorithm that predicts precisely sex-specific risks of ADE.

Galeano et al. [12] developed a machine learning framework for predicting side effects for drugs undergoing clinical development. Earlier, Benin et al. [13] reported the use of ML to successfully process the ADE reports in Connecticut Children’s Medical Center. Moreover, drug responses can be predicted using patients’ genetic information, enabling the identification of individuals at increased risk of adverse drug events (ADEs) and supporting decision-making regarding appropriate dosing and drug substitutions [14].

AI has the ability to transform PV through analyzing and processing large amounts of data that would normally be out of the scope of a human’s lifetime. The automation and ML models can optimize PV processes and provide a more efficient way to analyze information relevant to safety.

Storing large volumes of safety data electronically allows AI tools to significantly enhance data processing, reducing the burden of time, effort, errors, and costs. Moreover, eventually may lead to enhancement of clinical decisions by identifying ADEs patterns and trends.

A reliable AI/ML implementation must offer benefits that outweigh negative effects and ensure that unacceptable effects can be monitored for resolution. The growth of this field may be limited by challenges related to the lack of validated, established uses of AI in real-life safety settings. Ahire et al. [15] published a comprehensive detailed article about the potential advantages and challenges of integrating AI in the science of PV. AI-enhanced PV presents both potential and challenges, such as resolving data privacy issues and modifying regulatory frameworks. Shamim et al. [16] reviewed the current limitations of AI applications, regarding the dependency of AI model performance on the data quality and quantity, and the need of transparency and explainability to fill the gap between development and integration of AI model algorithms.

Several recommendations could be identified from the literature [15,17,18]. Although there is already much advancement made, there are still challenges and limitations that must be addressed to understand the complexities of these technologies.

Developing ICSRs and processing high volumes of data are major contributors to the stagnancy in technological advancement in the PV industry. AI models should go beyond correlation-based approaches by integrating causal inference techniques, which will allow for a more accurate understanding of the relationship between drugs and ADRs. Consideration is required for AI/ML systems to learn from previous analyzed data, while maintaining the capability to identify unexpected outcomes. The use of well-defined human understandable knowledge structures could increase the outcomes’ understanding, also playing a key role towards the adoption of AI-based systems.

Health organizations should play a significant role in regulating and implementing AI for PV and they should support emerging research and development of computational approaches and relevant data infrastructures. AI experts should appreciate the complexity of data interpretation and medical professionals should comprehend the AI technicality.

Table 1 list summarized points regarding the advantages, challenges and recommendation for the optimum integration of AI/ML technologies in the field of PV and medication safety.

The transformative process of integration of AI technologies in PV activities has revolutionizing effect on the performance of ADEs detection, monitoring and assessment. AI algorithms and automation have the advantages of saving time, reducing efforts and improving operational efficiency. There are recent evidences from the literature that support the applicability and benefits of AI and ML in PV field. Besides the significant benefits, barriers like data issues, ethical considerations, and interpretability are also exist. These issues can impact the performance of AI algorithms. Healthcare providers, regulators, and industry should exert more efforts to collaborate to overcome the challenges and to maximize potential applications of AI in public health and patient safety. Overall, the integration of AI in PV can lead to enhancement of drug surveillance and ultimately improve patient outcomes and regulatory decision making.