ARTIFICIAL INTELLIGENCE IN PHARMACY- A BRIDGE BETWEEN PAST AND FUTURE

ARTIFICIAL INTELLIGENCE IN PHARMACY- A BRIDGE BETWEEN PAST AND FUTURE

Mohd. Wasiullah 1, Piyush Yadav 2, Ankita singh3*

1. Principal, Department of Pharmacy, Prasad institute of technology, Jaunpur,UP,India
2. Head, Department of Pharma: Chemistry, Prasad institute of technology, Jaunpur, UP, India
3. Scholar, Department of Pharmacy, Prasad institute of technology, Jaunpur, UP, India

Abstract

By fusing conventional methods with cutting-edge, data-driven strategies, artificial intelligence (AI) is transforming the pharmacy industry. Pharmacy, which has historically relied on empirical knowledge, manual compounding, and established protocols, presents difficulties in handling complicated medicines, polypharmacy, and customized patient care. AI improves drug discovery, formulation optimization, pharmacokinetic/pharmacodynamic modeling, clinical decision support, and customized therapy using machine learning, deep learning, natural language processing, and predictive modeling. Predictive, adaptive, and patient-centered pharmaceutical care is made possible by emerging technologies like generative AI, digital twins, and smart pharmacy systems. Adoption of AI, despite its potential, necessitates resolving issues with data quality, interoperability, algorithmic bias, ethical responsibility, regulatory compliance, and workforce preparedness. A path for safer, more effective, and future-ready pharmacy is provided by integrating AI with traditional medicine systems, historical pharmacognosy, and human expertise. This will ultimately change the field from reactive to predictive, evidence-based care.

Keywords:Artificial intelligence, pharmacy, drug discovery, precision medicine, digital health, pharmacokinetics, pharmacovigilance, smart pharmacy.

Corresponding Author

Ankita singh, Scholar, 

Malhotra College of Pharmacy, Bhopal, Madhya Pradesh

Received: 29/12/2025                               

Revised: 12/01/2026                               

Accepted: 26/01/2026

DOI: No DOI

Copyright Information 

© 2026 The Authors. This article is published by Global Journal of Pharmaceutical and Scientific Research Copyright Author (s) 2024 Distributed under Creative Commons CC-BY 4.

How to Cite

Singh A, Wasiullah M, Yadav P. Artificial intelligence in pharmacy: A bridge between past and future. Global Journal of Pharmaceutical and Scientific Research. 2026; 2(1):108- 131. ISSN: 3108-0103.

1. Introduction: Redefining Pharmacy in the Age of Intelligence

From early apothecaries who compounded herbal cures to contemporary clinical pharmacies that prioritize patient-centered care and evidence-based therapies, pharmacy has undergone significant change over the course of centuries. Heuristic decision-making, manual preparation, and established clinical guidelines were the mainstays of pharmaceutical practice in the past. Although these methods guaranteed baseline safety and effectiveness, they were unable to handle the increasing complexity of contemporary treatments, such as polypharmacy, pharmacogenomic variability, and the management of chronic illnesses (Topol, 2019; Atanasov et al., 2021).

Artificial intelligence (AI) has emerged as a disruptive force in pharmacy thanks to the quick digitalization of healthcare data and increases in computing capacity. A wide range of methods, such as machine learning, deep learning, natural language processing, and network-based modeling, are included in artificial intelligence (AI). Together, these methods enable systems to learn patterns, make predictions, and assist in decision-making in ways that were previously unattainable (Liu et al., 2021; Rasmy et al., 2018). AI speeds up lead optimization, virtual screening, and target identification in drug discovery. AI-driven prediction models enhance stability, bioavailability, and manufacturability in formulation science. Clinically, AI enhances patient care by predicting adverse drug reactions, optimizing individualized dosing, and facilitating medication adherence (Walters & Murcko, 2020; Rieke et al., 2020).

Beyond efficiency, AI acts as a link between contemporary data-driven insights and conventional pharmaceutical expertise. AI-based virtual screening, bioactivity prediction, and network pharmacology are increasingly being combined with historical fields like pharmacognosy and herbal medicine, enabling the methodical analysis and validation of centuries' worth of empirical information (Chen et al., 2018). In a similar vein, AI uses real-world information gathered from wearable technology, claims data, and electronic health records to improve clinical judgment, spot warning signs, and guide precision medicine projects (Sherman et al., 2016).

AI integration in pharmacy poses particular difficulties despite its potential. For AI to reach its full potential, a number of crucial obstacles must be removed, including data quality, model transparency, ethical responsibility, regulatory compliance, and workforce preparedness (Floridi et al., 2018; Vayena et al., 2018). The clinical judgment, ethical thinking, and patient-centric attitude that characterize the pharmacy profession must be complemented by AI, which should be viewed as a collaborative tool.

Drug research and development, formulation optimization, clinical pharmacy, pharmacokinetics/pharmacodynamics, regulatory science, and future-ready pharmacy practice are just a few of the areas in which this paper examines the relationship between AI and pharmacy. This review highlights how the combination of human expertise and artificial intelligence can transform pharmacy into a predictive, efficient, and patient-centered discipline, bridging the lessons learned from the past with the innovations of the future. It does this by critically analyzing both the transformative potential and the challenges of AI.

2. Foundations of Artificial Intelligence for Pharmacy

Computational systems created to carry out tasks that normally require human intelligence, such as learning, reasoning, observation, and decision-making, are referred to as artificial intelligence (AI). AI has become a game-changing framework in pharmacy that makes it possible to analyze intricate, high-dimensional datasets produced by clinical practice, pharmaceutical development, drug discovery, and healthcare systems (Vamathevan et al., 2019). AI improves efficiency and accuracy throughout the pharmaceutical pipeline by enabling predictive modeling, pattern identification, and data-driven decision support through the integration of sophisticated algorithms with biological and clinical data.

2.1 Core Concepts of Artificial Intelligence

A variety of methodological techniques are included in artificial intelligence (AI), which aims to enable robots to simulate intelligent behavior. Learning paradigms including supervised learning, unsupervised learning, and reinforcement learning are fundamental to AI and have a direct bearing on the pharmaceutical sciences (Rajkomar et al., 2019).

In supervised learning, algorithms are trained using labeled datasets to predict preset outcomes, such as drug–target interactions, therapeutic response, or toxicity endpoints. In contrast, unsupervised learning finds latent structures in unlabeled data and is especially useful for biomarker discovery, disease subtyping, and patient stratification. Reinforcement learning has demonstrated potential in dosage optimization and adaptive clinical trial design. It focuses on sequential decision-making, where an agent learns optimal methods through reward-based feedback (Yu et al., 2018).

Feature representation, which entails transforming unprocessed pharmaceutical data—such as molecular fingerprints, physicochemical descriptors, omics data, or clinical variables—into relevant features appropriate for algorithmic learning, is a crucial part of developing AI models. In recent years, greater attention has been made on explainable artificial intelligence (XAI) to promote transparency, interpretability, and confidence in AI-driven choices, particularly in regulatory and therapeutic contexts (Samek et al., 2019).

2.2 Machine Learning and Deep Learning in Pharmaceutical Research

Algorithms may discover patterns from data without explicit programming thanks to machine learning (ML), a significant subfield of artificial intelligence. ML techniques have become crucial to pharmaceutical research due to their capacity to handle nonlinear interactions and big datasets. Quantitative structure–activity relationship (QSAR) modeling, pharmacokinetic property prediction, and evaluation of absorption, distribution, metabolism, excretion, and toxicity (ADMET) profiles all make extensive use of algorithms like random forests, support vector machines, and gradient boosting models (Vamathevan et al., 2019).

Multi-layered artificial neural networks that can automatically extract features and represent hierarchical data are used in deep learning (DL), a sophisticated subset of machine learning. Convolutional neural networks (CNNs) are commonly applied to molecular images and biomedical imaging, while recurrent neural networks (RNNs) and long short-term memory (LSTM) networks are suited for sequential and time-series data such as pharmacokinetic profiles and electronic health records (LeCun et al., 2015; Esteva et al., 2019). More recently, drug–target interactions and biological activity have been better predicted thanks to graph neural networks' (GNNs) greater performance in modeling chemical structures as graphs (Gaudelet et al., 2021).

The adoption of ML and DL has significantly accelerated early-stage drug discovery by enabling high-throughput virtual screening, reducing experimental attrition, and improving decision-making in lead optimization and candidate selection (Arora & Banerjee, 2022).

2.3 Natural Language Processing and Big Data Analytics in Healthcare

The goal of the specific AI field of natural language processing (NLP) is to enable computers to comprehend, analyze, and produce human language. NLP is crucial in pharmacy and healthcare to extract useful information from unstructured data sources, such as clinical narratives, pharmacovigilance reports, electronic health records, scholarly literature, and regulatory submissions (Kreimeyer et al., 2017). NLP approaches assist automatic detection of adverse drug reactions, drug–drug interactions, and prescription errors that may not be recorded through structured datasets alone.

By facilitating the integration and analysis of sizable, diverse datasets produced by genetics, proteomics, real-world evidence, wearable technology, and healthcare information systems, big data analytics enhances natural language processing. Real-time safety monitoring, individualized therapeutic optimization, and population-level risk prediction are all supported by AI-driven big data platforms (Ristevski & Chen, 2018). The use of NLP and big data analytics in pharmacovigilance has enhanced post-marketing drug surveillance and regulatory decision-making by increasing the speed and sensitivity of safety signal detection (Sarker et al., 2015).

A key component of AI-enabled pharmacy, NLP and big data analytics work together to promote the shift to precision medicine and enable evidence-based practice.

Table 1: Key Applications of Artificial Intelligence in Pharmacy

Figure 1: AI-Driven Pharmaceutical Workflow

3. The Pre-AI Era: Conventional Pharmacy Paradigms

Pharmacy practice and pharmaceutical sciences were mostly dependent on empirical knowledge, linear workflows, and human-driven decision-making prior to the integration of artificial intelligence. Major treatment advancements were made possible by these traditional paradigms, but they were frequently marked by expensive expenditures, drawn-out schedules, and a lack of flexibility in response to intricate biology and clinical heterogeneity. Understanding the constraints of the pre-AI age is vital to grasp the revolutionary role of AI in modern pharmacy.

3.1 Traditional Drug Discovery and Development Models

Target identification was the first step in the sequential, hypothesis-driven process of traditional drug discovery, which was followed by hit discovery, lead optimization, preclinical testing, and multi-phase clinical trials. This pipeline was heavily dependent on trial-and-error experimentation, high-throughput wet-lab screening, and incremental optimization tactics (Hughes et al., 2011). The overall success rate of converting candidate compounds into authorized medications remained low despite technical advancements.

According to DiMasi et al. (2016), the conventional strategy is linked to lengthy development delays, frequently surpassing 10 to 15 years, and a significant financial burden, with estimates indicating an average cost of more than USD 2 billion per authorized medicine. Late-stage clinical trials have especially high attrition rates, usually as a result of ineffectiveness or unexpected toxicity that previous models were unable to predict (Waring et al., 2015).

Additionally, conventional in vitro and animal models frequently showed poor translatability due to their limited predictive validity for human pharmacological reactions. Innovation and efficiency in drug development were hampered by the lack of sophisticated computational tools, which limited the capacity to examine intricate molecular interactions, biological networks, and illness heterogeneity.

3.2 Limitations in Clinical Decision-Making and Patient Care

In the pre-AI period, clinical pharmacy decision-making was largely dependent on standardized treatment guidelines, professional experience, and population-averaged clinical trial data. While evidence-based medicine considerably improved service quality, it often failed to account for interindividual variability in genetics, comorbidities, and drug response (Topol, 2019). As a result, therapeutic decisions were typically reactive rather than predictive.

Medication mistakes, drug-drug interactions, and inappropriate dose were more likely when medication therapy management relied on manual examination of patient data, especially in polypharmacy and chronic disease care (Bates et al., 1998). Pharmacovigilance operations were similarly confined, as adverse drug reaction (ADR) detection depended heavily on spontaneous reporting systems, which are known to suffer from underreporting and delayed signal detection (Hazell & Shakir, 2006).

Additionally, the increasing amount of patient data and biomedical literature surpassed human cognitive capacity, making it more difficult for doctors to effectively synthesize knowledge. These limitations brought to light the necessity of intelligent systems that can integrate various data sources to facilitate individualized, real-time clinical decision-making.

3.3 Manufacturing, Quality Control, and Supply Chain Challenges

In the past, batch-based production methods were used in conventional pharmaceutical manufacturing, and end-product testing rather than ongoing monitoring was the main method used to ensure quality. According to Rathore and Winkle (2009), this reactive strategy frequently led to process unpredictability, material waste, and delayed discovery of quality problems. Proactive process control and optimization were hampered by the restricted application of real-time data analytics.

The majority of quality control procedures were labor-intensive and manual, depending on offline analytical testing and predetermined requirements. Batch failure and regulatory non-compliance were made more likely by these systems' inability to quickly adjust to process changes or raw material variability (FDA, 2004).

Prior to artificial intelligence, supply chain management had to deal with issues including logistical coordination, inventory optimization, and demand forecasting. Due to their frequent fragmentation and lack of data, these systems were susceptible to interruptions, shortages, and inefficiencies (Shah, 2004). The capacity to foresee dangers and react quickly to market and healthcare demands was hampered by the lack of predictive analytics.

4. AI-Driven Drug Discovery and Development

Artificial intelligence has radically changed drug discovery and development by enabling data-driven decision-making across several phases of the pharmaceutical pipeline. AI-driven methods, in contrast to conventional linear workflows, combine molecular, biological, and clinical data to expedite target selection, optimize lead compounds, lower attrition rates, and shorten development schedules (Vamathevan et al., 2019). The long-standing issues with cost, inefficiency, and poor translatability that defined traditional drug discovery paradigms are addressed by these technologies.

4.1 Intelligent Target Identification and Validation

A crucial first stage in the drug development process is target identification, which entails choosing biological molecules whose modification may have therapeutic benefits. Traditionally, this procedure depended on limited experimental data and hypothesis-driven approaches, often resulting to inefficient target selection. Large-scale omics datasets, protein–protein interaction networks, and disease-specific molecular signatures may now be systematically analyzed using AI-based techniques to find new and clinically relevant therapeutic targets (Zhang et al., 2022).

In order to identify disease-associated pathways and rank targets according to anticipated druggability and safety profiles, machine learning algorithms can combine genomic, transcriptomic, proteomic, and phenotypic data (Hopkins & Groom, 2002; Costa et al., 2020). By analyzing target centrality, redundancy, and possible off-target impacts within biological networks, network-based and systems biology approaches further improve target validation. These AI-powered tactics lower the chance of late-stage clinical failure and increase target confidence.

4.2 Virtual Screening and Predictive Lead Optimization

A computational method for finding promising compounds from vast chemical libraries is called virtual screening. Because AI can accurately predict binding affinity, selectivity, and biological activity, it has greatly improved both structure-based and ligand-based virtual screening (Gomes et al., 2017). By capturing the nonlinear interactions between chemical structure and biological function, deep learning models—such as convolutional and graph neural networks—perform better than conventional docking techniques.

AI-driven predictive models direct repeated chemical changes in lead optimization to enhance pharmacokinetic characteristics, potency, and selectivity while reducing toxicity (Yang et al., 2019). These methods make it possible to prioritize high-quality candidates early in the discovery phase and lessen reliance on expensive experimental assays. Consequently, AI-driven virtual screening has emerged as a crucial instrument for expediting the hit-to-lead and lead optimization phases.

4.3 De Novo Drug Design and Molecular Modeling

The process of creating new chemical entities with desired pharmacological properties is known as de novo drug design. The automated creation of chemically valid, biologically active compounds has been made possible by recent developments in artificial intelligence, especially generative models like variational autoencoders (VAEs), generative adversarial networks (GANs), and reinforcement learning frameworks (Zhavoronkov et al., 2019).

AI-driven molecular modeling provides simultaneous optimization of numerous objectives, including potency, selectivity, solubility, and synthesis feasibility. Large chemical regions that are unreachable by conventional medicinal chemistry methods can be quickly explored by these models (Schneider & Clark, 2019). Integration of AI with molecular dynamics simulations significantly improves structure-based drug design by enabling accurate prediction of ligand–target interactions and conformational flexibility.

4.4 AI-Based ADMET and Toxicity Prediction

One of the main reasons for attrition in drug development is still adverse absorption, distribution, metabolism, excretion, and toxicity (ADMET) failure. AI-based predictive models have become effective instruments for early pharmacokinetic and safety profile evaluation, greatly lowering late-stage failures (Pires et al., 2015).

Oral bioavailability, blood–brain barrier permeability, metabolic stability, cardiotoxicity, and hepatotoxicity may all be precisely predicted by machine learning and deep learning algorithms trained on extensive experimental datasets (Wu et al., 2018). AI-driven ADMET prediction improves overall development efficiency by prioritizing safer drug candidates and facilitating informed decision-making by early detection of potential safety risks.

4.5 Artificial Intelligence in Drug Repurposing

Finding novel therapeutic uses for already-approved medications, or "drug repurposing," provides a quick and affordable substitute for conventional drug development. By combining chemical, biological, clinical, and real-world data to find new drug–disease connections, AI has significantly improved repurposing efforts (Pushpakom et al., 2019).

To identify repurposing opportunities with high translational potential, machine learning algorithms examine pathway disruptions, gene expression profiles, and electronic health records (Xue et al., 2018). During the COVID-19 pandemic, AI-driven drug repurposing became very popular as computational screening quickly found potential treatments for testing and clinical assessment. These methods show how AI can speed up the development of new treatments, especially for uncommon and neglected illnesses.

Table 2: AI-Driven Drug Discovery vs Conventional Drug Discovery

Figure 2: Comparison of Conventional vs AI-Driven Drug Discovery

5. Smart Formulation Development and Intelligent Manufacturing

Pharmaceutical formulation and manufacturing have typically depended on empirical experimentation and rigid process controls. Although these methods guaranteed adherence to regulations, they frequently lacked adaptability, effectiveness, and predictive power. A shift toward data-driven design, real-time process optimization, and proactive quality assurance has been made possible by the incorporation of artificial intelligence (AI) into formulation science and manufacturing. This is in line with contemporary regulatory requirements for pharmaceutical development that is risk- and science-based (Rathore & Winkle, 2009; Yu et al., 2014).

5.1 Machine Learning in Pharmaceutical Formulation Optimization

Excipient composition, processing conditions, and dosage form characteristics are only a few of the interdependent variables that must be optimized during formulation creation. Traditional trial-and-error methods need a lot of time and resources. For determining nonlinear correlations between formulation factors and critical quality attributes (CQAs) including dissolving rate, stability, and bioavailability, machine learning (ML) models have become effective tools (Aksu et al., 2012).

Solid dosage forms, nanoparticle-based systems, and controlled-release formulations have all been effectively optimized using algorithms like artificial neural networks, random forests, and support vector machines (Agatonovic-Kustrin & Beresford, 2000). Rapid formulation space screening, less experimental work, and increased final product resilience are all made possible by ML-driven formulation optimization. These methods are especially useful for complicated delivery systems where formulation variables show high complexity and interdependence, such as biologics and nanoformulations.

5.2 Artificial Intelligence in Quality by Design (QbD)

The methodical approach to pharmaceutical development known as Quality by Design (QbD) places a strong emphasis on comprehending and managing unpredictability in order to guarantee predetermined product quality. By facilitating sophisticated modeling of the connections between critical material attributes (CMAs), critical process parameters (CPPs), and CQAs, AI improves QbD (Yu et al., 2014).

Machine learning models improve risk assessment, design space exploration, and process optimization by learning from historical production data and experimental outcomes (Santos et al., 2019). Proactive quality assurance as opposed to reactive quality control is made possible by AI-supported QbD frameworks, which enhance decision-making during formulation development and scale-up. AI-assisted QbD is becoming more widely acknowledged by regulatory bodies as a way to improve pharmaceutical product lifecycle management and process comprehension.

5.3 Process Analytical Technology and Real-Time Process Control

Through timely assessments of crucial quality and performance qualities, Process Analytical Technology (PAT) seeks to design, evaluate, and regulate pharmaceutical manufacturing processes. PAT tools were frequently restricted to offline or near-line analysis in conventional systems. The integration of AI with PAT has enabled real-time monitoring, predictive control, and adaptive manufacturing (FDA, 2004).

In order to identify deviations, forecast process outcomes, and suggest corrective steps prior to quality failures, AI-driven PAT systems integrate sensor data, spectroscopy, and sophisticated analytics (Fonteyne et al., 2015). These features improve process robustness, lower batch failures, and reduce waste while supporting continuous manufacturing models. AI-enabled PAT is increasingly essential to next-generation pharmaceutical production as continuous manufacturing receives regulatory approval.

5.4 Predictive Maintenance and Supply Chain Intelligence

Because pharmaceutical production supply chains and infrastructure are intricate and heavily regulated, they are susceptible to interruptions and inefficiencies. In order to detect failures, optimize maintenance schedules, and minimize unscheduled downtime, AI-based predictive maintenance systems evaluate equipment performance data (Lee et al., 2018). This proactive strategy reduces maintenance costs while increasing operational dependability and compliance.

By combining real-time data from production, distribution, and market consumption, AI-driven analytics in supply chain management improve demand forecasting, inventory optimization, and logistics coordination (Ivanov et al., 2019). These clever methods guarantee timely medication availability, lessen drug shortages, and strengthen the resilience of the supply chain. The necessity of AI-enabled supply chain intelligence in preserving pharmaceutical continuity during worldwide disruptions was further brought to light by the COVID-19 pandemic.

Table 3: AI Applications Across the Pharmaceutical Lifecycle

6. Transforming Clinical Pharmacy and Patient Care

Optimizing prescription use through pharmacist knowledge and guideline-based decision-making has been the traditional focus of clinical pharmacy. However, traditional care models have been put to the test by growing illness complexity, polypharmacy, and the amount of patient-specific data. By providing data-driven decision support, tailored therapy, proactive medication management, and improved drug safety surveillance, artificial intelligence (AI) is revolutionizing clinical pharmacy and eventually enhancing patient safety and therapeutic outcomes (Topol, 2019).

6.1 AI-Enabled Clinical Decision Support Systems

One of the first and most significant uses of AI in healthcare is clinical decision support systems (CDSS). To help medical professionals make well-informed treatment decisions, AI-enabled CDSS incorporates patient data, test findings, medication histories, and clinical recommendations. AI-driven CDSS, in contrast to rule-based systems, continuously learns from real-world data, increasing accuracy and adaptability over time (Rajkomar et al., 2019).

AI-powered CDSS supports dose modification, drug-drug interaction detection, contraindication identification, and therapy optimization in difficult instances including critical care and renal impairment in clinical pharmacy practice (Kawamoto et al., 2005). These systems have showed decreases in drug mistakes and improvements in adherence to evidence-based prescribing, demonstrating their significance in promoting patient safety and clinical efficiency.

6.2 Precision Medicine and Personalized Therapeutic Strategies

The goal of precision medicine is to customize treatment plans according to each patient's unique genetic, environmental, and lifestyle features. By combining genomic data with clinical and phenotypic data to forecast drug response and maximize therapy selection, artificial intelligence (AI) plays a key role in this paradigm (Collins & Varmus, 2015).

Pharmacogenomics has effectively used machine learning algorithms to find genetic variations linked to therapeutic efficacy and adverse effects, allowing for customized drug selection and dosage (Relling & Evans, 2015). By matching patients to medicines that are most likely to be beneficial while avoiding toxicity, AI-driven precision medicine techniques have improved treatment stratification and outcomes in oncology and chronic illness management (Esteva et al., 2019).

6.3 Intelligent Medication Therapy Management and Adherence

Drug therapy management (MTM) has historically relied on patient counseling and manual drug regimen revision. By integrating digital health technology, pharmacy dispensing data, and electronic health records, AI improves MTM by enabling ongoing monitoring of drug use, adherence patterns, and clinical outcomes (Krumme et al., 2012).

According to Choudhry et al. (2014), AI-driven systems can forecast therapy-related issues, identify patients who are at high risk of non-adherence, and create customized intervention methods. By offering real-time feedback and reminders, wearable technology and mobile health apps enhance AI-based adherence monitoring and improve long-term treatment outcomes, especially in chronic conditions like diabetes and cardiovascular disorders.

6.4 Pharmacovigilance and Adverse Drug Reaction Monitoring

Pharmacovigilance, which has historically relied on manual case evaluation and spontaneous reporting systems, is an essential part of patient safety. These methods are constrained by inadequate data, underreporting, and reporting delays (Hazell & Shakir, 2006). By making it possible to automatically identify adverse drug reactions (ADRs) from a variety of data sources, such as social media platforms, clinical narratives, and electronic health records, artificial intelligence (AI) has revolutionized pharmacovigilance (Sarker et al., 2015).

By detecting previously undiscovered drug-event connections and ranking safety signals for regulatory assessment, natural language processing and machine learning algorithms improve signal identification (Inácio & Cavaco, 2019). AI-driven pharmacovigilance systems promote safer and more effective drug usage by facilitating earlier intervention, better risk management, and more effective post-marketing surveillance.

7. Artificial Intelligence in Pharmacokinetics and Pharmacodynamics

By characterizing drug disposition and drug-response linkages, pharmacokinetics (PK) and pharmacodynamics (PD) provide the quantitative basis for sensible pharmacological therapy. Although robust, compartmental models and nonlinear mixed-effects methods used in conventional PK/PD modeling may have trouble capturing nonlinearities, high-dimensional variables, and intricate biological interactions. Increased prediction accuracy, customized dosing, and system-level comprehension of drug action are all made possible by artificial intelligence (AI), which has become a potent addition to traditional PK/PD frameworks (McComb et al., 2022).

7.1 AI-Based PK/PD Modeling and Simulation

AI-based PK/PD modeling learns intricate correlations between drug exposure, biological factors, and pharmacological responses using machine learning (ML) and deep learning approaches. AI-driven methods, in contrast to conventional mechanistic models, can manage large-scale, diverse datasets without necessitating stringent model structural assumptions (Zhang et al., 2020).

Concentration–time profiles, exposure–response correlations, and interindividual variability have all been accurately predicted using neural networks, random forests, and Gaussian process models (Brennan et al., 2018). Because they maintain physiological interpretability while enhancing predictive accuracy, hybrid modeling approaches—which combine mechanistic PK/PD principles with AI algorithms—are becoming more and more popular (Holford et al., 2021). During the early and late stages of drug development, these AI-enhanced simulations aid in scenario testing, dose optimization, and decision-making.

7.2 Population Pharmacokinetics and Dose Individualization

The goal of population pharmacokinetics (PopPK) is to measure drug exposure variability among patient populations and find factors that affect drug disposition. AI has expanded PopPK by enabling automated covariate selection, nonlinear association identification, and real-time dose individualization (Scotton et al., 2021).

Based on demographics, organ function, comorbidities, and concurrent drugs, machine learning models trained on clinical and real-world data can forecast patient-specific PK characteristics (Fuchs et al., 2018). In particular populations where typical PopPK models may be constrained, such as children, geriatrics, and critically ill patients, these models have shown enhanced performance (Woillard et al., 2020). Precision dosing techniques are supported by AI-driven dose individualization, which maximizes therapeutic efficacy while reducing toxicity.

7.3 Systems Pharmacology and Network-Based Drug Action Analysis

Systems pharmacology combines systems biology with PK/PD modeling to clarify pharmacological effects at the molecular, cellular, and organismal levels. By identifying important regulatory nodes, signaling pathways, and feedback mechanisms involved in medication action, artificial intelligence (AI) makes it possible to analyze complicated biological networks (Zhao & Iyengar, 2012).

In order to forecast pharmacodynamic results, off-target effects, and synergistic drug combinations, network-based machine learning techniques model drug–target–pathway interactions (Cheng et al., 2019). In multifactorial diseases like cancer, neurological conditions, and metabolic syndromes, where therapeutic effects go beyond specific targets, these techniques are especially helpful. AI-driven systems pharmacology supports rational combination therapy design and translational research while improving mechanistic understanding.

8. AI in Clinical Trials and Regulatory Science

Among the most resource-intensive stages of drug development are clinical trials and regulatory evaluation, which are frequently hampered by high failure rates, protracted schedules, and intricate compliance requirements. Artificial intelligence has become a game-changing instrument to boost evidence production, increase trial efficiency, and facilitate data-driven regulatory decision-making (Unger et al., 2020).

8.1 AI-Optimized Clinical Trial Design and Patient Recruitment

By enhancing protocol viability, endpoint selection, and statistical power, AI-driven analytics are being employed more and more to optimize clinical trial design. To forecast trial results and spot design flaws early in development, machine learning algorithms examine past trial data, disease registries, and real-world datasets (Harrer et al., 2019).

In clinical trials, patient recruitment continues to be a significant bottleneck, with many studies falling short of enrollment goals. AI-based screening systems combine electronic health records (EHRs), genomic data, and NLP-extracted clinical notes to match eligible patients to trials more effectively (Fogel, 2018). These strategies have shown increases in population diversity, retention rates, and recruitment speed while lowering operating expenses (Liu et al., 2021).

8.2 Real-World Evidence Generation and Data Integration

In regulatory science and post-marketing assessment, real-world evidence (RWE)—which comes from sources like wearable technology, claims databases, registries, and electronic health records—has grown in significance. By spotting trends, confounders, and treatment effects that might not be recorded in controlled trials, AI makes it easier to integrate and analyze heterogeneous real-world data (Sherman et al., 2016).

In order to extract clinically significant insights from unstructured data, such as physician notes and patient-reported outcomes, natural language processing is essential (Wang et al., 2020). In addition to conventional randomized controlled trials, AI-powered RWE analytics facilitate label expansion, comparative effectiveness studies, and long-term safety monitoring.

8.3 Regulatory Decision Support and Compliance Automation

AI-enabled solutions are being investigated by regulatory bodies more and more to assist with risk assessment, compliance monitoring, and decision-making. AI systems can assist in assessing big regulatory submissions by automating data validation, spotting discrepancies, and prioritizing crucial safety signals (Topol, 2019).

AI-driven systems improve quality audits, pharmacovigilance reporting, and adherence to Good Manufacturing Practice (GMP) and Good Clinical Practice (GCP) regulations from a compliance standpoint (Bate & Hobbiger, 2021). Transparency, explainability, and strong validation are necessary for regulatory adoption of AI, but these technologies are anticipated to be crucial to the future regulatory ecosystem.

9. Bridging Past Knowledge with Future Pharmacy

In addition to being a symbol of technical progress, AI acts as a link between contemporary data-driven innovation and conventional pharmacological understanding. A more comprehensive and sustainable pharmacy paradigm is made possible by combining AI with traditional methods, natural product research, and human expertise.

9.1 Integration of AI with Pharmacognosy and Natural Products Research

Historically, pharmacognosy has depended on labor-intensive screening of natural compounds and empirical information. By enabling high-throughput virtual screening, bioactivity prediction, and structure–activity relationship modeling of phytochemicals and natural substances, artificial intelligence has revived this research (Chen et al., 2018).

Lead identification can be accelerated while maintaining existing knowledge systems by using machine learning models trained on phytochemical databases and bioassay data to predict the therapeutic potential, toxicity, and biological targets of natural compounds (Atanasov et al., 2021). The scientific validity and widespread acceptance of plant-based therapies are improved by this combination.

9.2 Digitization and AI-Driven Analysis of Traditional Medicine Systems

Due to a lack of digitization and standardization, traditional medical systems like Ayurveda, Traditional Chinese Medicine (TCM), and Unani are frequently neglected while being rich reservoirs of therapeutic information. The methodical examination of conventional formulations, herb-herb interactions, and multi-target processes is made possible by AI-driven data mining and network pharmacology techniques (Li et al., 2017).

AI analytics and the digitization of historic literature and clinical records make it easier to create evidence, assess safety, and incorporate traditional medicine into contemporary healthcare systems (Patwardhan et al., 2015). This confluence promotes pharmacy practices that are both scientifically sound and inclusive of all cultures.

9.3 Human–AI Collaboration in Modern Pharmacy Practice

AI is a supplementary tool that improves professional decision-making, not a substitute for human competence, despite its transformative potential. In contemporary pharmacy practice, human-AI collaboration allows pharmacists to concentrate on ethical decision-making, patient counseling, and clinical judgment while AI manages data-intensive activities (Bates et al., 2020).

Clear accountability structures, transdisciplinary training, and faith in AI systems are necessary for successful implementation. AI promotes safer, more effective, and patient-centered pharmacy practice by enhancing human intellect rather than replacing it.

10. Ethical, Legal, and Societal Considerations

Questions of accountability, justice, and societal effect are inextricably linked to the ethical integration of AI in pharmacy. Ethical foresight is crucial rather than optional since AI systems, in contrast to traditional decision-support tools, may independently affect clinical decisions, regulatory results, and medication access (Vayena et al., 2018).

10.1 Data Privacy, Security, and Governance

Pharmaceutical AI systems rely on constant access to sensitive resources, such as genomic profiles, medication histories, electronic health records, and empirical data. Concerns about patient permission, the use of secondary data, and international data exchange are increased by these data sources (Price & Cohen, 2019). Misuse, re-identification hazards, and a decline in public confidence can result from inadequate governance.

Lifecycle-based data stewardship, where privacy protections extend from data capture through model deployment and post-market surveillance, is emphasized by modern governance systems. To balance large-scale analytics with patient confidentiality, strategies like federated learning, safe multiparty computation, and differential privacy are being suggested more and more (Rieke et al., 2020). Strong governance is essential for pharmacies to maintain long-term social acceptability of AI-driven healthcare innovation as well as legal compliance.

10.2 Algorithmic Bias, Transparency, and Explainability

Incomplete clinical representation, unbalanced datasets, or historical injustices ingrained in healthcare service are common causes of bias in AI systems. Biased models may negatively impact drug dosage, therapeutic eligibility, or safety forecasts in underrepresented populations in pharmacy settings (Rajkomar et al., 2018).

Therefore, two essential ethical and legal standards are transparency and explainability. Pharmacists and regulators can comprehend, question, and interpret algorithmic outputs using explainable AI techniques as feature attribution, surrogate modeling, and counterfactual explanations (Samek et al., 2021). Explainability is particularly critical for high-stakes decisions such as dose individualization, adverse event prediction, and regulatory clearance support, where opaque models may erode accountability.

10.3 Ethical Responsibility and Trust in AI-Assisted Pharmacy

Trust in AI-assisted pharmacy arises from a mix of technical reliability, ethical governance, and professional monitoring. Regardless of algorithmic autonomy, human stakeholders such as pharmacists, physicians, developers, and regulators continue to bear a strong ethical duty (London, 2019).

As a bridge between patients and AI systems, pharmacists hold a special ethical position. Their responsibilities include ensuring that human values—such as empathy, autonomy, and beneficence—are upheld, evaluating AI recommendations, and contextualizing outputs within patient-specific conditions. Clear communication with patients regarding the function and constraints of AI in therapeutic decision-making is necessary to build trust.

11. Emerging Trends and Future Perspectives

The integration of systems biology, digital health infrastructure, and artificial intelligence will shape pharmacy in the future. Pharmaceutical systems are becoming more anticipatory, adaptive, and constantly learning, according to emerging trends.

11.1 Generative AI and Autonomous Drug Discovery

A completely new paradigm in drug development has been brought about by generative AI, which makes it possible to algorithmically create unique chemical structures with predetermined pharmacological restrictions. In contrast to conventional virtual screening, generative models actively create molecules while concurrently optimizing for synthetic feasibility, potency, and selectivity (Walters & Murcko, 2020).

A future where AI systems iteratively design, test, and refine drug concepts with little human interaction is suggested by the steady rise of semi-autonomous discovery platforms. Although this poses concerns about intellectual property rights and validation, generative AI has revolutionary potential to solve unmet medical needs and spur innovation.

11.2 Digital Twins and Predictive Therapeutics

AI, pharmacometrics, and real-time clinical data come together to form digital twins, which allow for virtual testing on customized patient models. Digital twins can be used in pharmacy to model drug-drug interactions, disease development, and PK/PD behavior prior to actual administration (Björnsson et al., 2020).

Proactive intervention tactics are made possible by digital twin-supported predictive therapies, which minimize bad effects and reduce trial-and-error prescribing. Digital twins may also help with lifecycle drug review, dose optimization research, and in silico trials when regulatory environments change.

11.3 Smart Pharmacies and an AI-Ready Pharmaceutical Workforce

Smart pharmacies combine robots, clinical intelligence systems, and AI-powered automation to improve patient-centered treatment and operational effectiveness. AI-enabled pharmacies facilitate population-level drug optimization, tailored counseling, and predictive inventory management in addition to dispensing automation (Kassam et al., 2021).

However, workforce transformation must keep pace with technology innovation. Future pharmacists will need to be proficient in transdisciplinary teamwork, AI ethics, and data analysis. To guarantee that pharmacists continue to be leaders in AI-driven healthcare settings rather than merely passive users, education and ongoing training are crucial.

12. Challenges and Limitations of AI in Pharmacy

Although use of AI in pharmacy has advanced quickly, there are still issues that could limit scalability, dependability, and equity if they are not resolved.

12.1 Data Quality, Interoperability, and Standardization Issues

The reliability of AI systems depends on the quality of the training data. Datasets in pharmacy are frequently dispersed within commercial, clinical, and regulatory silos, with different forms and varying quality. Cross-institutional validation is hampered and model generalizability is limited by a lack of interoperability (Hripcsak et al., 2015).

Robust AI implementation requires standardized ontologies, harmonized data models, and strict curation procedures. Even sophisticated algorithms may produce untrustworthy or deceptive results if fundamental data issues are not resolved.

12.2 Model Validation, Interpretability, and Clinical Translation

The discrepancy between retrospective model performance and actual clinical utility is a significant barrier to AI in pharmacy. While many AI systems show remarkable performance in controlled datasets, they are unable to adjust to changing clinical settings (Sendak et al., 2020).

Prospective validation, ongoing observation, and compliance with legal requirements are necessary for clinical translation. Whether or not AI models can be trusted and included into standard pharmacological decision-making still depends critically on their interpretability.

12.3 Infrastructure, Cost, and Skill-Gap Barriers

AI technology implementation necessitates large investments in safe computer environments, digital infrastructure, and qualified human resources. These regulations may worsen global disparities in pharmaceutical innovation by disproportionately harming low- and middle-income environments (Bohr & Memarzadeh, 2020).

Coordinated efforts combining academia, business, regulators, and professional associations are necessary to close infrastructure and skill gaps. Ensuring equitable access to AI-enabled pharmacy innovations requires inclusive training programs and scalable, open-source solutions.

13. Conclusion

Artificial intelligence is changing every aspect of pharmacy, from clinical decision-making and patient care to drug development and formulation. AI bridges the knowledge of old procedures with contemporary computational insights, enabling quicker, more accurate, and customized pharmacological interventions through the integration of machine learning, deep learning, natural language processing, and predictive modeling. New developments that show promise for predictive, adaptive, and patient-centered care include generative AI, digital twins, and smart pharmacy systems. However, issues including data quality, interoperability, algorithmic bias, ethical responsibility, regulatory compliance, and workforce readiness must be resolved for deployment to be successful. Human-AI collaboration is the key to the future of pharmacy, as pharmacists use AI to improve clinical judgment, guarantee safety, and maximize therapeutic results. Pharmacy can transform from reactive and generalized practices into a predictive, effective, and patient-centered discipline—truly bridging the lessons of the past with the innovations of the future—by responsibly integrating AI with historical pharmacological knowledge, conventional medical systems, and empirical data.

14. Acknowledgement

The authors sincerely acknowledge the support of their institution and colleagues who provided valuable insights during the preparation of this review.

15. Conflict of Interest

The authors declare that there are no conflicts of interest associated with this work.

16. References

• Agatonovic-Kustrin, S., & Beresford, R. (2000). Basic concepts of artificial neural network (ANN) modeling and its application in pharmaceutical research. Journal of Pharmaceutical and Biomedical Analysis, 22(5), 717–727. https://doi.org/10.1016/S0731-7085(99)00272-1
• Aksu, B., De Beer, T., Folestad, S., Ketolainen, J., Lindén, H., Lopes, J. A., Sandler, N., & Yliruusi, J. (2012). Strategic approaches to process analytical technology (PAT) implementation and process control in pharmaceutical manufacturing. International Journal of Pharmaceutics, 417(1–2), 1–15. https://doi.org/10.1016/j.ijpharm.2011.06.048
• Arora, N., & Banerjee, A. K. (2022). Artificial intelligence in drug discovery: Opportunities, challenges, and future prospects. Drug Discovery Today, 27(4), 1035–1043. https://doi.org/10.1016/j.drudis.2021.11.019
• Atanasov, A. G., Zotchev, S. B., Dirsch, V. M., & Supuran, C. T. (2021). Natural products in drug discovery: Advances and opportunities. Nature Reviews Drug Discovery, 20(3), 200–216. https://doi.org/10.1038/s41573-020-00114-z
• Bate, A., & Hobbiger, S. F. (2021). Artificial intelligence, real-world automation and the safety of medicines. Drug Safety, 44(2), 125–132. https://doi.org/10.1007/s40264-020-01010-8
• Bates, D. W., Levine, D., Syrowatka, A., Kuznetsova, M., Craig, K. J. T., Rui, A., Jackson, G. P., Rhee, K., & Bates, D. W. (2020). The potential of artificial intelligence to improve patient safety: A scoping review. NPJ Digital Medicine, 3, 54. https://doi.org/10.1038/s41746-020-0263-6
• Bates, D. W., Cullen, D. J., Laird, N., Petersen, L. A., Small, S. D., Servi, D., … Leape, L. L. (1998). Incidence of adverse drug events and potential adverse drug events: Implications for prevention. JAMA, 274(1), 29–34. https://doi.org/10.1001/jama.274.1.29
• Björnsson, B., Borrebaeck, C., Elander, N., Gasslander, T., Gawel, D. R., Gustafsson, M., … Benson, M. (2020). Digital twins to personalize medicine. Genome Medicine, 12, 4. https://doi.org/10.1186/s13073-019-0709-0
• Bohr, A., & Memarzadeh, K. (2020). The rise of artificial intelligence in healthcare applications. In Artificial Intelligence in Healthcare (pp. 25–60). https://doi.org/10.1016/B978-0-12-818438-7.00002-2
• Cheng, F., Kovács, I. A., & Barabási, A. L. (2019). Network-based prediction of drug combinations. Nature Communications, 10(1), 1197. https://doi.org/10.1038/s41467-019-09186-x
• Chen, Y., de Bruyn Kops, C., & Kirchmair, J. (2018). Data resources for the computer-guided discovery of bioactive natural products. Journal of Chemical Information and Modeling, 57(9), 2099–2111. https://doi.org/10.1021/acs.jcim.7b00341
• Collins, F. S., & Varmus, H. (2015). A new initiative on precision medicine. New England Journal of Medicine, 372(9), 793–795. https://doi.org/10.1056/NEJMp1500523
• DiMasi, J. A., Grabowski, H. G., & Hansen, R. W. (2016). Innovation in the pharmaceutical industry: New estimates of R&D costs. Journal of Health Economics, 47, 20–33. https://doi.org/10.1016/j.jhealeco.2016.01.012
• Esteva, A., Robicquet, A., Ramsundar, B., Kuleshov, V., DePristo, M., Chou, K., … Dean, J. (2019). A guide to deep learning in healthcare. Nature Medicine, 25(1), 24–29. https://doi.org/10.1038/s41591-018-0316-z
• Fogel, D. B. (2018). Factors associated with clinical trials that fail and opportunities for improving the likelihood of success: A review. Contemporary Clinical Trials Communications, 11, 156–164. https://doi.org/10.1016/j.conctc.2018.08.001
• Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., … Vayena, E. (2018). AI4People—An ethical framework for a good AI society. Minds and Machines, 28(4), 689–707. https://doi.org/10.1007/s11023-018-9482-5
• Gaudelet, T., Day, B., Jamasb, A. R., Soman, J., Regep, C., Liu, G., … Bronstein, M. M. (2021). Utilising graph machine learning within drug discovery and development. Briefings in Bioinformatics, 22(6), bbab159. https://doi.org/10.1093/bib/bbab159
• Hansen, L. K., & Müller, K. R. (2021). Explainable AI: Interpreting, explaining and visualizing deep learning. Springer.
• Hazell, L., & Shakir, S. A. W. (2006). Under-reporting of adverse drug reactions: A systematic review. Drug Safety, 29(5), 385–396. https://doi.org/10.2165/00002018-200629050-00003
• Hopkins, A. L., & Groom, C. R. (2002). The druggable genome. Nature Reviews Drug Discovery, 1(9), 727–730. https://doi.org/10.1038/nrd892
• Holford, N., Ma, G., & Ploeger, B. A. (2021). Clinical trial simulation: A review. Clinical Pharmacology & Therapeutics, 109(2), 360–372. https://doi.org/10.1002/cpt.2001
• Hughes, J. P., Rees, S., Kalindjian, S. B., & Philpott, K. L. (2011). Principles of early drug discovery. British Journal of Pharmacology, 162(6), 1239–1249. https://doi.org/10.1111/j.1476-5381.2010.01127.x
• Kassam, R., Collins, J. B., Berkowitz, J., & Bushell, M. (2021). Artificial intelligence and the future of pharmacy practice. Canadian Pharmacists Journal, 154(4), 205–213. https://doi.org/10.1177/17151635211013545
• Kreimeyer, K., Foster, M., Pandey, A., Arya, N., Halford, G., Jones, S. F., … Botsis, T. (2017). Natural language processing systems for capturing and standardizing unstructured clinical information: A systematic review. Journal of Biomedical Informatics, 73, 14–29. https://doi.org/10.1016/j.jbi.2017.07.012
• LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444. https://doi.org/10.1038/nature14539
• Rajkomar, A., Dean, J., & Kohane, I. (2019). Machine learning in medicine. New England Journal of Medicine, 380(14), 1347–1358. https://doi.org/10.1056/NEJMra1814259
• Rasmy, L., Xiang, Y., Xie, Z., Tao, C., Zhi, D., & Tao, C. (2018). A study of deep learning approaches for medication recommendation using EHRs. Journal of Biomedical Informatics, 85, 91–101. https://doi.org/10.1016/j.jbi.2018.07.004
• Ristevski, B., & Chen, M. (2018). Big data analytics in medicine and healthcare. Journal of Integrative Bioinformatics, 15(3), 20170030. https://doi.org/10.1515/jib-2017-0030
• Topol, E. J. (2019). High-performance medicine: The convergence of human and artificial intelligence. Nature Medicine, 25(1), 44–56. https://doi.org/10.1038/s41591-018-0300-7
• Vamathevan, J., Clark, D., Czodrowski, P., Dunham, I., Ferran, E., Lee, G., … Zhao, S. (2019). Applications of machine learning in drug discovery and development. Nature Reviews Drug Discovery, 18(6), 463–477. https://doi.org/10.1038/s41573-019-0024-5
• Walters, W. P., & Murcko, M. A. (2020). Assessing the impact of generative AI on medicinal chemistry. Nature Biotechnology, 38(2), 143–145. https://doi.org/10.1038/s41587-020-0418-2
• Woillard, J. B., Labriffe, M., Debord, J., & Marquet, P. (2020). Machine learning for therapeutic drug monitoring: A systematic review. Clinical Pharmacokinetics, 59(12), 1451–1466. https://doi.org/10.1007/s40262-020-00908-3
• Zhao, S., & Iyengar, R. (2012). Systems pharmacology: Network analysis to identify multiscale mechanisms of drug action. Annual Review of Pharmacology and Toxicology, 52, 505–521. https://doi.org/10.1146/annurev-pharmtox-010611-134520
• Zhavoronkov, A., Ivanenkov, Y. A., Aliper, A., Veselov, M. S., Aladinskiy, V. A., Aladinskaya, A. V., … Aspuru-Guzik, A. (2019). Deep learning enables rapid identification of potent DDR1 kinase inhibitors. Nature Biotechnology, 37(9), 1038–1040. https://doi.org/10.1038/s41587-019-0224-x