Exploring Transformative Trends: A Thorough Examination Of The Role Of Artificial Intelligence In Healthcare And Pharmaceutical Research & Development

Artificial Intelligence (AI) is a branch of science that focuses on intelligent machine learning, particularly smart computer programs that deliver outcomes resembling the human thought process. Typically, this process involves acquiring. Creating streamlined systems for utilizing data, drawing clear or tentative conclusions, and making self-corrections or adjustments. AI is typically employed to analyze machine learning in order to replicate the cognitive tasks performed by humans. AI technology is utilized to conduct precise analyses and gain valuable insights(1). In AI technology, a blend of helpful statistical models and computational intelligence is utilized. The advancement and creativity in AI applications are frequently linked with concerns about the potential threat of unemployment. Nevertheless, nearly all progressions in. The versatile applications of AI technology are widely praised for the confidence it brings, significantly enhancing its effectiveness across industries. AI technology has recently become an essential component of various industries due to its valuable applications in numerous technical and research sectors. The progressive move towards incorporating AI technology in pharmacy, such as in drug discovery, drug delivery, and formulation development, is gaining momentum. Other healthcare applications have transitioned smoothly from being hyped to offering hopeful prospects. AI models are utilized to forecast in vivo responses, pharmacokinetic parameters, appropriate dosages, and other aspects related to therapeutics(2). In light of its significance. The utilization of in silico models in pharmacokinetic drug prediction enhances the efficiency and cost-effectiveness of drug research. There exist two fundamental categories of advancements in AI technology. The initial category encompasses traditional computing approaches such as expert systems, which have the ability to mimic human experiences and clarify conclusions based on principles, such as expert systems. The second category includes systems that can simulate brain functioning using artificial neural networks (ANNs). Particularly, different types of artificial neural networks such as deep neural networks (DNNs) and recurrent neural networks (RNNs) play a key role in shaping the advancements of AI technology. Participating in the Merck Kaggle and NIH Tox21 challenge, DNN showcased its capabilities effectively. Concerns demonstrate a higher level of predictability compared to the standard machine learning techniques. Machine learning utilizes appropriate statistical methods that enable learning either with or without explicit programming. Moreover, de novo design fosters the creation of novel drug molecules. In terms of the best or preferred characteristics. The current review delves into the diverse applications of AI in the field of pharmacy, pinpointing drug discovery, drug delivery formulation development, poly-pharmacology, and hospital pharmacy(3).

Terminologies used in Artificial Intelligence

• Neural network: Neural network algorithms are organized in a sequence that aligns beforehand to recognize fundamental patterns within a dataset. The ongoing process replicates the functioning of the human brain. The neural networks operate on a system of interconnected neurons in the aforementioned process. Artificial elements and possibly those of a natural origin can coexist. Neural networks can adapt to different inputs, leading to optimal outcomes. The network was developed without the necessity of revising the output criteria for its design. The idea of neural networks, stemming from artificial intelligence, is rapidly becoming popular in the development of trading systems.
• Basics of Neural Networks: In the realm of industrial evolution, financial circumstances play a crucial role when it comes to neural networks, enabling processes like evolution to thrive. Algorithmic trading, forecasting time-series, and classifying securities. A neural network has the capability to function in a similar and intelligent manner. Similar to the neural network found in the human brain. Neurons in a neural network engage in mathematical function processing to gather and categorize information based on specific patterns. In a neural network, nodes are linked together in layers, creating an interconnected structure. Every node represents a perceptron, which is an algorithm utilized for. Supervised learning of binary classifiers is akin to conducting a multiple linear regression analysis. Perceptrons are arranged systematically. Interconnected layers are present in a multi-layered perceptron, known as MLP. Input patterns are collected by the input layers. The output layer consists of classifications or output signals that input patterns can correspond to. The weightings of the input are delicately adjusted by the hidden layers until the neural network reaches its minimum margin of error. When considering the outputs, it is assumed that the hidden layers capture significant features from the input data that are predictive. The achievement of a utility resembling statistical methods like principal component analysis involves extracting features from the outline(4).
• Fuzzy Logic: Fuzzy logic is widely employed in expert systems as a component of modern control systems. Fuzzy logic possesses the capacity to solve problems similarly to humans, but at a much faster pace than the human mind, particularly in the context of decision-making. The process involves gathering information and transforming it into more meaningful insights by organizing it into limited truths known as Fuzzy sets. In big corporate settings, business uses it to enhance decision-making support systems and personal growth. It is advantageous in the chemicals manufacturing sector as it helps to regulate the pH levels of chemicals, aids in drying processes, and facilitates chemical distillation. Fuzzy logic is gaining popularity in the field of Artificial Intelligence due to its applications in natural language processing. Fuzzy Logic is employed alongside Neural Networks to mimic decision-making processes more quickly and effortlessly than the human brain. Success is achieved through gathering data, which is then transformed into more meaningful information by creating partial truths known as Fuzzy sets(5).
• NeuroFuzzy logic: Fuzzy systems utilize learning algorithms derived from neural network theory to regulate their parameters through data sample manipulation. Modern neuro-fuzzy systems are typically represented as specialized multilayer feed-forward neural networks, for example. ANFIS, GARIC, NEFCLASS, and so forth. Here is a notable illustration of fuzzification within neural network design: Self-organizing feature maps. In neuro fuzzy networks, the propagation process, connection weights, and activation functions exhibit variations compared to those in traditional neural networks. Various approaches have been identified through collected data, with the term most commonly used in the field of neuro-fuzzy systems being the one that exhibits distinct properties. The theory of neural networks assists fuzzy systems by training them using algorithms provided by neural network theory. Local information is utilized to guide the learning process and also plays a role in local adjustments within the basic fuzzy system. A three-layer feed forward neural network represents the essence of a neurofuzzy system. The input variables are depicted in the initial layer, the fuzzy rules in the middle layer, also called the hidden layer, and the output variables in the final and third layer. The process involves representing fuzzy connection weights as encoding of fuzzy sets. It's not necessary to represent a fuzzy system in this manner in order to apply a learning algorithm to it(6).
• Evolutionary computing: It is a commonly used term that describes a computational process where solutions evolve through the application of inheritance rules. The processes of recombination, mutation, and selection. In the realm of research focusing on formulation, evolutionary algorithms stand out as a noteworthy subset. Playing a significant role in the field of formulation research is its application.
• Genetic algorithms: In 1970, John Holland employed Genetic algorithms. A search method ideally tailored for optimization purposes; based on the assumption of experimentation. Genetic algorithms offer an iterative process that drives the evolution of the population. The solution's initial population is created and established. The fitness of each population member is assessed as part of this process(7).
• Machine learning (ML): Machine Learning, or ML, is a technique utilized for analyzing data by automatically constructing analytical models through algorithms that continuously learn from data or valuable insights. Machine learning is categorized into three primary classes: supervised learning, unsupervised learning, and reinforcement learning. Further categorizing these three classes, we have supervised learning, which branches into classification for disease diagnosis and regression for drug reinforcement/ADMET prediction. Entered without supervision. The clustering process leads to the discovery of disease subtypes, while feature finding leads to the discovery of disease targets. Reiteration Gaining knowledge leads to making decisions that ultimately result in the creation of new drug designs and the planning of experimental designs(8).

Current Pharmaceutical Challenges and the Role of AI

Smoothly put, continuous research is being conducted in the pharmaceutical field on small molecules in order to enhance product quality and meet customer needs, as they offer various benefits. The chemical synthesis procedure is uncomplicated, and the production of synthetic derivatives is cost-effective. Numerous reliable and efficacious small-molecule-loaded formulations can be found in the pharmaceutical industry. Apart from the targeted therapies for uncommon illnesses, numerous groundbreaking small molecules encounter competition. Complex data and clinical trials are necessary for the launch of generic molecules. These processes serve to amplify the financial strain on businesses, prompting their involvement. Further advancement in creativity and new ideas(9). Nevertheless, the biomolecular drug sector continues to expand rapidly to address the challenges caused by limitations in the small molecular size and inadequate spread of research and advancements. The behaviour of small molecules is influenced by their shape and how they interact chemically. Biomolecules, comprised mainly of amino acids from protein sources, also incorporate nucleotides or ribonucleotides for nucleic acids, being large molecular structures. The stability and function of these are further shaped by the supramolecular sequence and spatial conformation. Certain biomolecules, like insulin and adalimumab, have proven to be highly effective products(10). The pharmacokinetics of these molecules are intricate, with infusion being the preferred and most practical administration route for these biomolecules. Pharmacokinetic modulation and molecular stabilization play crucial roles in nucleic acid-based research. Achieving optimal pharmacokinetic exposure and enhancement of these molecular forms are essential objectives to pursue. The latest technological advancements could offer valuable assistance in tackling these challenges and resolving associated issues. While there is significant potential for AI in advancing drug delivery techniques and drug discovery, it does exhibit certain noteworthy constraints that necessitate human intervention or expertise to decipher the intricate outcomes(11). The significant advancements in AI forecasting rely on datasets, yet the understanding of the outcomes, due to the ambiguous nature, necessitates human intervention for arriving at the proper conclusion. Artificial intelligence may encounter challenges related to algorithm bias when processing information to make predictions and evaluate hypotheses. In addition, it is quite common for docking simulations to yield certain outcomes(12).

In the exploration of molecules that are not currently active. Hence, a thorough examination of these factors continues to rely on human engagement in order to make informed decisions and conduct cross-checks to eliminate any potential system biases. However, the realm of AI holds vast possibilities for diverse applications. Through diligent efforts, it is conceivable that the constraints linked to AI can be alleviated, paving the way for its efficacy and dependability. In the realm of AI, the approach utilized includes the application of machine learning, including its components like deep learning and natural language processing. The learning process can take on different forms, either under supervision or independently, with the choice of algorithm used playing a vital role. In supervised learning, we utilize familiar inputs (features) and corresponding outputs (labels or targets), unlike unsupervised learning that focuses on unfamiliar outputs. The supervised approach entails predicting the output, which can be labels or targets, by considering various inputs or features. Conversely, unsupervised classification strives to form clusters that exhibit homogeneity within them in terms of characteristics(13).

Supervised AI Learning: Supervised learning involves training an algorithm on a labeled dataset, where the desired output is already known. The algorithm acquires the skill of linking input data to the precise output through the examination of patterns and connections. Among the data that has been labeled. This methodology is frequently employed across a range of applications, including image recognition, natural language processing, and predictive modeling. Task-oriented strategies entail establishing particular objectives to attain desired results from a specific input. This method involves using labeled data to train algorithms for tasks like data classification or outcome forecasting. The main tasks in supervised learning typically revolve around classification, where the goal is to predict a label, and regression, where the aim is to predict a quantity. Several methods can be utilized to tackle supervised learning tasks, depending on the situation. The characteristics of the data within a specific problem field. Some of the techniques encompassed are Naïve Bayes, K-nearest neighbors, support vector machines, ensemble learning, random forest, and linear models. Regression, support vector regression, and additional techniques(14). It has several applications in the pharmaceutical industry, as described below:

• Drug Discovery and Design: In the realm of Drug Discovery and Design, supervised learning algorithms offer a powerful tool that enables the prediction of the activity or properties of potential new drug candidates. Through training on a dataset containing identified compounds and their corresponding activities, the model is able to grasp intricate patterns and connections between molecular attributes and intended results. This facilitates the anticipation of the effects, strength, or potential harmfulness of new substances, which is helpful in the process of discovering and creating medications(15).
• Predictive Maintenance and Quality Control: Supervised learning in pharmaceutical manufacturing is commonly applied for tasks such as predictive maintenance and ensuring quality control. Through training with data obtained from manufacturing processes, equipment sensor data, or quality testing results, the model can become adept at predicting equipment failure, product quality deviations, or process abnormalities. This enables proactive maintenance and quality assurance measures to be implemented(16).
• Drug Target Identification: Supervised learning methods have the capacity to assist in recognizing potential. Analyze biological data to identify drug targets. Training on data containing genetic, proteomic, or transcriptomic features along with their association to drug response or disease progression enables the model to learn patterns and recognize potential outcomes. Potential areas for further exploration.
• Disease Diagnosis and Prognosis: Supervised learning models have the potential to be utilized for diagnostic purposes. Predicting patient outcomes or determining diseases based on medical data. By training on labeled datasets that include patient attributes, clinical data, and disease outcomes, the model is able to classify patients into separate disease categories, predict disease progression, and estimate treatment responses.
• Adverse Event Detection: Supervised learning algorithms are capable of being utilized on pharmacovigilance data for the purpose of detecting and categorizing adverse events linked to medications. Through training on labeled adverse event reports, the model is able to grasp patterns and pinpoint potential safety signals, aiding in the detection process. Analysis of unfavorable incidents(17).
• Predictive Modeling for Clinical Trials: Supervised learning is a valuable tool for forecasting results in clinical trials. By utilizing past clinical trial data features such as patient attributes, treatment approaches, and trial results, the model has the ability to predict patient response, treatment effectiveness, and safety outcomes. This data has the potential to inform trial planning and enhance the process of selecting suitable patients.

Unsupervised AI Learning: Unsupervised learning delves into a branch of machine learning in which the algorithm operates without the aid of labeled data. Rather, its role is to discern patterns and connections within the data autonomously. This method is frequently employed in exploratory data analysis and can prove helpful in uncovering concealed patterns or groupings within a dataset. The method being discussed is often referred to as a "data-driven approach," commonly recognized in various fields. The goal is to uncover patterns, structures, or insights from data that has not been annotated. Numerous common unsupervised tasks exist, such as clustering, dimensionality reduction, and visualization. Discovering association rules and detecting anomalies. Many unsupervised learning tasks can be tackled by utilizing widely-used techniques like clustering algorithms such as hierarchical clustering. Various algorithms such as clustering (including K-means, K-medoids, single linkage, complete linkage, BOTS), association learning, and feature selection and extraction techniques (such as Pearson correlation, principal component analysis) are utilized based on the data's unique characteristics(18). AI's unsupervised learning techniques can prove beneficial for pharmaceutical applications, specifically for tasks like exploratory analysis, pattern recognition, and data visualization, as outlined below:

• Clustering: Clustering algorithms organize data points by their similarities, enabling the detection of inherent groupings or clusters within the data. Clustering is commonly utilized in pharmaceutical settings to group different types of data, including gene expression profiles, chemical structures, and patient information, with the aim of identifying subgroups that share similar characteristics. This can assist in identifying specific targets, categorizing patients, and distinguishing different types of compounds or illnesses.
• Dimensionality Reduction: Reduction techniques like principal component analysis can be employed to streamline the data analysis process.
• Principal Component Analysis (PCA) and t-Distributed Stochastic Neighbor Embedding (t-SNE) are commonly employed to simplify high-dimensional datasets while retaining valuable information. These methods are beneficial for visualizing and examining complex concepts. Analyze datasets to determine primary variables or features that aid in enhancing decision-making procedures. Dimensionality reduction techniques can be used on a variety of pharmaceutical data types, such as gene expression data, drug activity profiles, or imaging data.
• Anomaly Detection: Anomaly detection algorithms pinpoint uncommon or exceptional data. Points that show significant deviation from the anticipated patterns. Within the pharmaceutical sector, anomaly detection proves beneficial for uncovering adverse events, pinpointing potential safety issues, and revealing data quality concerns. Anomalous behaviour without supervision. Detection techniques like the local outlier factor (LOF) or isolation forest can be useful in identifying unusual patterns or data points that may require closer examination.
• Association Rule Mining: Association rule mining techniques, such as the Apriori algorithm, are utilized to uncover intriguing relationships or associations among items within a dataset. In the pharmaceutical sector, association rule mining is employed to investigate drug interactions, adverse event information, and trends of co-occurrence among medical conditions and medications. These techniques offer valuable insights into potential drug interactions, medication patterns, and support pharmacovigilance efforts.
• Topic Modelling: Topic modelling algorithms, like latent Dirichlet allocation (LDA), uncover hidden topics or themes within extensive text datasets. In the pharmaceutical sector, topic modelling offers a valuable tool for examining scientific literature, clinical trial reports, and social media data to uncover crucial research themes, evolving patterns, and patient perspectives. This can be beneficial for literature mining, competitive intelligence, or gaining insight into patient viewpoints(19, 20).

AI Tool Application in Dosage form Designs

To better understand how medications are distributed and their impact. The human body is compartmentalized. Biological membranes are of utmost importance in this particular situation. The permeation rate is influenced by the pathway. It is crucial to monitor the administration in order to evaluate performance effectively. A well-designed medication delivery system. Passive diffusion is reliant on. Ongoing research is delving into the chemical properties of the medication and its active components. Penetration primarily starts with membrane transport. Depends on complex biological processes. To investigate the pharmacokinetic characteristics of the compound. A newer computer model has been created for drug delivery systems. The text has been submitted for your kind consideration. The predictability of preclinical models defines their reliability. It is paramount to establish specific parameters in the pharmaceutical industry. Discovery and creativity in the field. Artificial Intelligence possesses the ability to provide cutting-edge solutions. Technology is applied to analyze complex sets of data, resulting in important findings. Improving comprehension of research entities can be enhanced by utilizing AI. To carry out pharmacokinetic research utilizing state-of-the-art AI technology. Artificial neural networks are also employed in the utilization of technologies. Offer databases such as chemical, genomic, and phenotypic. Improve your understanding of drug interactions by making use of databases. The sophisticated operational mechanisms of molecules. Certain people might struggle to comprehend the topic. Various techniques are utilized to explore the effects of the. Research is currently being conducted on how the drug delivery method affects pharmacokinetics. Medication facilitates a more profound comprehension to be acquired. Disposition and toxicity are the topics at hand. Artificial intelligence has the capability to assist in drug development and repurposing strategies, such as utilizing existing treatments for new conditions. Consideration of patients’ needs and medical conditions plays a significant role in the formulation, pharmacokinetics, and development of medications. One of the main challenges hindering the widespread implementation of AI in building delivery systems is the limited availability of accurate information databases. By intelligently using available information, analyzing large amounts of data, and enhancing the knowledge database through successful self-supervised experiments linked to precise parameter tracking, AI supports upcoming applications. To sum up, artificial intelligence plays a crucial role in the field of drug delivery by tackling issues such as forecasting drug distribution, comprehending drug interactions, and designing efficient drug delivery mechanisms. Through the use of AI tools and databases, we can enhance the efficiency and effectiveness of drug delivery systems across diverse clinical environments(21, 22).

AI for Drug Delivery

The integration of AI and big data in the pharmaceutical sector has led to the development of computational pharmaceutics, aimed at enhancing drug delivery systems through the application of multiscale modeling techniques. AI algorithms along with machine learning techniques play a crucial role in computational pharmaceutics by analyzing vast datasets. This enables researchers to predict drug behavior, evaluate multiple scenarios, and enhance drug delivery systems efficiently, eliminating the need for time-consuming trial-and-error studies. This helps to reduce the time required for medication development, cut costs, and increase efficiency. Through the use of computational pharmaceutics, researchers are able to delve into the interactions between drugs and the human body, thus paving the way for more personalized and efficient treatment alternatives(23). Furthermore, by incorporating AI algorithms into drug delivery systems, we can effectively pinpoint potential side effects or interactions between medications, thus bolstering patient safety and minimizing the chances of adverse reactions. AI algorithms are able to examine intricate relationships among drug properties, formulation elements, and physiological factors in order to predict drug performance across various levels, ultimately enhancing our understanding of drug delivery mechanisms and aiding in the creation of efficient drug delivery systems. It assists in forecasting the physicochemical properties of the medication, its in vitro drug release behavior, and stability. The same approach is also employed to enhance in vivo pharmacokinetic parameters, drug distribution, and research on in vivo-in vitro correlation. Researchers can identify potential risks and challenges related to drug delivery systems at an early stage in the development process by leveraging the right mix of AI tools. This enables them to make proactive refinements and modifications, thereby reducing risks and enhancing the efficiency of drug delivery systems(24).

AI for oral solid dosage form development

Tablets stand out as one of the most widely employed solid dose forms within the pharmaceutical industry. AI has the capability to assist in finding the best formulations and exploring favorable process attributes. Artificial neural networks (ANNs), fuzzy logics, neural networks, and evolutionary algorithms are utilized to create reliable dosage forms and enhance understanding of processing and operational inputs and outputs. Tablets are widely favored as a solid form of medication, constituting a significant portion of the pharmaceutical delivery sector. This product is crafted by blending active pharmaceutical ingredients with excipients, then gently shaping and sizing them through crushing or molding processes(25). Excipients are added to tablets to influence specific product outcomes, such as aiding in tablet disintegration, dissolution, and the release of medication. Artificial intelligence can also be applied to predict drug release and analyze the effects of key tablet production factors in the realm of systemic drug delivery, ensuring reliable quality control processes. Some AI programs have been utilized for identifying imperfections in tablets. Being able to forecast drug release through formulations has the potential to offer dependable quality control. Experiments on drug release are carried out through both in vivo and in vitro methods, seen as fundamental technologies consistently evaluated during the progress of product development(26). Artificial intelligence is utilized in the formulation of medications to assist in predicting drug release, leading to a reduction in the number of necessary iterations to optimize the batch. This in turn minimizes the effort and expenses associated with pilot batch scaling and production processes. Artificial intelligence could be utilized to forecast drug release and dissolution patterns, along with investigating disintegration timing to help select the optimal batch for further processing at a larger scale(27).

AI for nanomedicine

Leveraging artificial intelligence (AI) abilities in data analysis, pattern recognition, and optimization holds significant promise for researchers in the field of nanomedicine. This integration holds the promise of speeding up the exploration of innovative nanoscale treatments, enhancing diagnostics, refining drug administration, and propelling personalized medicine forward. The remarkable impact of AI in nanomedicine lies in its capacity to deliver targeted and precise treatment methods on a minuscule scale, offering the potential to transform the field of healthcare. Nanoparticles play a crucial role in the field of medicine delivery, where AI algorithms assist researchers in designing and enhancing nanoparticles tailored for specific purposes through the prediction of their physical and chemical properties, stability, and effectiveness(28). In the effective utilization of nanomedicines for delivering drugs, particularly in cancer treatment, AI-powered assistance is essential in tasks like choosing the right medication, determining dosage, and selecting stimuli-responsive materials. Specifically, deep learning algorithms employed in the treatment of melanoma have demonstrated remarkable precision in patient care and diagnostic procedures. AI algorithms play a crucial role in replicating the behavior and interactions of nanoscale materials within biological systems. The ability to model various aspects, such as nanoparticle behavior, drug release kinetics, and potential toxicity, paves the way for predicting outcomes. This, in turn, facilitates the creation of safe and efficient formulations for nanomedicine. Additionally, artificial intelligence is utilized in nanosensors and biosensors for real-time monitoring of biomarkers, medication levels, and the progression of illnesses(29). These sensors provide healthcare professionals with ongoing feedback, enabling prompt adjustments in treatment and personalized therapy modifications. Integrating AI-powered databases proves advantageous for enhancing the scalability of nanocarriers through automated processes. Artificial intelligence is applied in nanocarrier drug delivery systems to enhance the efficiency of nanocarriers and conduct compatibility testing for medications through computational approaches. The methodologies assess drug loading, formulation stability, and drug retention to enhance therapeutic nanocarriers for targeting specific cell types in tumor treatment. Researchers like Yuan He et al. utilize machine learning techniques in their studies. By using computational methods like Monte Carlo simulations and molecular dynamics, it is possible to predict the creation of nanocrystals through high-pressure homogenization and wet ball milling, thus bypassing the need for repetitive experiments(30). By collaborating with structural nanobiology and integrating artificial intelligence into the development of database repositories for nanocarriers, we enhance the ability to define 3D structures and investigate their physical and chemical properties. Exploring the correlation between nanocarrier structure and toxicological, physical, and biological data heavily relies on these repositories. Lutz Nuhn's research showcases how AI can be used to analyze the diverse vascular permeability of medication delivery systems based on nanoparticles. These discoveries will contribute to the advancement of protein nanoparticle drug delivery systems that possess active transendothelial permeability for targeting tumors. Likewise, Zhoumeng Lin and colleagues. Utilize artificial intelligence alongside physiologically based pharmacokinetic (PBPK) modeling to study cancer drugs, gaining a better understanding of the factors contributing to the reduced effectiveness of nanoparticle delivery to tumors(31).

Applicability of AI in parenteral, transdermal, and mucosal routes

Artificial intelligence (AI) is essential in the advancement and production of intricate pharmaceutical formulations, such as injectables and biologics. AI systems assist in forecasting physicochemical parameters for drug formulations, refining pH, solubility, stability, and viscosity through the examination of formulation components and manufacturing processes. This leads to the development of stable parenteral formulations with improved quality, efficiency, and variability. AI algorithms are essential in improving the production of parenteral products by revealing the critical process variables that influence product qualities. The immediate analysis of process data improves product consistency, reduces batch failures, and increases overall manufacturing efficiency. By examining large datasets obtained from analytical evaluations such as particle size analysis and chromatography, AI recognizes trends and discrepancies. This allows for the prompt identification and correction of quality issues, resulting in a more effective and dependable production process(32). In the realm of visual inspection, AI is used to examine the behavior of particles inside containers, distinguishing among swimming, sinking, and adhering particles. Methods such as particle tracking and deep learning algorithms are employed to ensure precise inspection and separation of particles. AI-driven image processing is similarly implemented to tackle challenges linked to parenteral batch defects, especially the formation of bubbles. AI also influences maintenance practices for intricate manufacturing machinery. AI algorithms anticipate equipment failures and schedule preemptive maintenance by evaluating sensor data, the history of equipment performance, and maintenance documentation. This not only removes downtime but also boosts productivity and reduces maintenance costs. AI also guarantees adherence to regulations in the production of parenteral and complex biological products. AI algorithms evaluate compliance by analyzing process data and product characteristics, identify potential noncompliance issues, and suggest enhancements to the process. This aids in meeting Good Manufacturing Practice (GMP) and regulatory compliance standards(33). In a significant case, Bannigan et al. emphasize the promise of machine learning (ML) technologies in the pharmaceutical sector, pointing out ML’s ability to accelerate the advancement of drug delivery technologies. In spite of the difficulties linked to neural network models, tree-based models such as LGBM demonstrate potential in reducing the time and expenses related to the development of long-acting injectable (LAI) formulations. Computational pharmaceutics, encompassing machine learning and multiscale simulations, has transformed conventional trial-and-error techniques in drug formulation for ocular, transdermal, pulmonary, and mucosal delivery systems. Recent progress in molecular simulations, mathematical modeling, and pharmacokinetic/pharmacodynamic (PK/PD) modeling accelerates product development. In silico modeling and simulations offer comprehensive insights, enabling a logical approach to formulation design and addressing data challenges via interdisciplinary collaborations, especially in the context of Pharma 4.0(34).

AI for Pharmacokinetics and Pharmacodynamics

The process of developing medication is intricate, encompassing phases such as drug discovery, preclinical studies, clinical trials, and regulatory endorsement. Pharmacokinetics and pharmacodynamics, which establish the optimal dosage, method of delivery, and safety of a medication in the body, are essential to this procedure. Conventional experimental methods for assessing these variables can be lengthy and expensive, and they often do not yield dependable predictions regarding therapeutic efficacy and safety(35). Historically, research in pharmacokinetics and pharmacodynamics relied on experimental methods like animal testing and human clinical trials, which posed challenges like ethical dilemmas, limitations in sample sizes, and variations among individuals. To address these issues, computer models and artificial intelligence (AI) technologies have emerged as effective tools for predicting drug pharmacokinetics and pharmacodynamics, offering a quicker, more economical, and precise method. AI has demonstrated considerable potential in the fields of pharmacokinetics, pharmacodynamics, and drug discovery. AI has been beneficial in forecasting and enhancing medication pharmacokinetics and pharmacodynamics as powerful computing and machine learning techniques have progressed. In spite of challenges such as large datasets and data reliability, AI exhibits capability for advancing PKPD research and its impact on therapeutics(36).

AI in Disease Diagnosis

Ensuring a considerate treatment design and patient wellbeing relies on the crucial element of disease analysis. Mistakes made by humans create obstacles to accurate diagnosis, and the misreading of information adds to the difficulty of this task. AI appears as a flexible answer, providing assurance in precision and effectiveness. An extensive literature review uncovers a range of technologies and methods utilized for disease diagnosis. The changing human population and diverse environmental factors consistently spur the increasing demand for the healthcare system(37). Despite weaknesses, inconsistencies, and unexamined discrepancies, recent findings indicate the creation of new techniques to clarify their relevance, highlighting previously uncharted situations. Classifying patients according to the severity of their illness emphasizes the significance of AI in the diagnostic procedure. Diagnosis, described as the condition assigned based on existing issues, highlights the necessity of keeping thorough health report forms for every patient. This approach assists in gathering feedback received through assessments and tests, facilitating prompt and customized healthcare requirements. Clinician-driven analysis is influenced by personal judgment and can differ, underscoring the necessity for diagnostic approaches that emphasize early disease identification. AI plays an essential part in detecting and determining the early predictive phase of illnesses, especially in instances of cancer and dementia(38). Existing technological methods utilize deep learning, neural networks, and algorithms for the identification, extraction, and processing of data. Remarkably, AI’s impact is profound in conditions like cancer and dementia, where impartial algorithms, supported by pertinent datasets, aid in making precise predictions. Deep learning connections adjust to evolutionary developments and improve forecasts, while unsupervised learning supports hepatitis diagnosis. Even though larger datasets and a wider range of inputs enhance AI functionality, the outcomes may be challenging to comprehend. Deep learning uses in diagnostics encompass the classification of skin conditions and the detection of atrial fibrillation. Cross-validation is a method for assessing algorithms that includes randomly dividing into various sets(39). The performance of AI is evaluated based on its accuracy, sensitivity, and specificity, with these three essential components highlighted in its usual evaluations. After examining the literature, the clinical elements managing deep learning networks and neural pathways, using support vector machines, nearest neighbours, random forests, decision trees, logistic regression, naive Bayes, discriminant analysis, and convolutional neural networks, can yield more thorough outcomes. Performance analysis based on algorithms may be conducted, factoring in elements like origin, sample size, and the quantity of features in training and testing samples. The integration of decision trees and reasoning was utilized for identifying liver diseases. Numerous studies on predictive modeling have been carried out, particularly in forecasting early Parkinson’s disease. Within the domain of lung disorders, a rib segmentation algorithm utilizing chest X-ray images was created to address the shortcomings of conventional rib-wise segmentation techniques. The algorithm delivers outstanding results, especially in rib segmentation, due to unpaired sample augmentation and a multi-scale network, and shows potential for use in detecting lung cancer and other respiratory diseases. In a recent study, algorithms and machine learning techniques have been employed to recognize and categorize cardiac arrhythmias through the analysis of ECG data. Another research implemented an optimization genetic algorithm (GA) along with a support vector machine (SVM) classifier to effectively classify and diagnose tuberculosis (TB)(40).

AI in Digital Therapy

AI possesses the ability to form significant connections within raw datasheets, enabling its application in illness diagnosis, treatment, and mitigation. A variety of innovative methodologies employed for computational comprehension in this growing field have the ability to be utilized across an extensive array of medical science disciplines. Addressing intricate clinical challenges requires substantial knowledge acquisition, analysis, and implementation. The advancement of medical AI supports healthcare professionals in the resolution of intricate clinical scenarios. Artificial neural networks, evolutionary computational models, fuzzy expert systems, and hybrid intelligent systems serve as valuable resources for healthcare workers concerning data storage and processing(41). The ANN functions based on the principles of a biological nervous system, utilizing a network of interconnected computer processors referred to as neurons that are capable of conducting parallel calculations for data processing. The initial artificial neuron was generated employing a binary threshold function. The multilayer feed-forward perceptron is a commonly utilized model that consists of three layers: an input layer, a hidden layer, and an output layer. Each neuron is interconnected through connections that have numerical weights. Paul Werbos introduced "Backpropagation learning," a distinctive method featuring an appropriately tailored learning algorithm, in 1974. The artificial neural network (ANN) has been utilized in image assessment, data analysis, and waveform examination. Fuzzy logic, which is a category of reasoning and inference, acknowledges and utilizes real-life situations by implementing a continuous spectrum of membership values from 0 to 1, representing falsehood and truth(42). Fuzzy controllers have been implemented in the surgical setting to administer vasodilators and anesthetics. The concept of "survival of the fittest" is highlighted in this evolutionary computing approach, which simulates the natural evolution process. The genetic algorithm, a widely used technique within this framework, explores a vast array of random solutions to a particular problem and ultimately identifies the optimal solution by discarding less effective options. Artificial intelligence also plays a crucial role in the realm of radiation. A novel technological advancement, automated treatment planning, has proven to be highly advantageous in the planning of radiation treatments. This automated system greatly enhances the quality of plans, their consistency, and reduces the rates of errors. There are three categories of treatment workflows: automated rule application, logic modeling based on historical clinical data, and multi-criteria optimization. Clinical guidelines can be applied through an easy-to-use computer program with set frameworks, enabling the effective assessment of patient anatomy and physiology. This automated therapy planning system has the ability to replicate the thought process involved in manual therapy planning. In particular, three-dimensional dose distribution and spatial dosage models exhibit promising precision. Radiomics, which utilizes various imaging biomarkers, provides comprehensive data regarding tumors. This method can be employed to forecast outcomes and measure toxicity in each radiation treatment patient(43).

Application of AI for 3D-Printed Dosage Forms

The utilization of AI in the realm of 3D-printed dosage forms has transformed pharmaceutical manufacturing by allowing for personalized medicine and improving drug delivery systems. AI algorithms have the capability to enhance the design and formulation of 3D-printed dosage forms according to patient-specific characteristics, including age, weight, and medical background, resulting in customized drug therapies. By utilizing machine learning and computational modeling, AI can evaluate extensive datasets and mimic the performance of 3D-printed dosage forms, facilitating the swift prototyping and refinement of drug release profiles, dosage amounts, and shapes(44). AI additionally assists in forecasting and addressing possible manufacturing obstacles, fine-tuning printing parameters, and guaranteeing quality assurance. Moreover, AI-infused feedback mechanisms can perpetually enhance the 3D-printing procedure by adapting from real-time data, improving precision, repeatability, and scalability(45). In summary, the use of AI in 3D-printed dosage forms offers significant promise in progressing personalized medicine and enhancing patient results. The tablets produced through 3D printing are created by employing the fused-filament method of manufacturing, along with binder jetting, laser sintering, and a pressure microsyringe. Important processing factors that influence the 3D-printed tablets include the temperature of both the nozzle and the platform as well as the printing speed. Obeid et al. illustrated how the processing parameters affect a 3D-printed tablet containing diazepam and its subsequent study on drug release using an ANN model. They investigated the infill pattern, infill density, and other input factors essential for effective drug dissolution in 3D-printed tablets. The relationships among the various variables were assessed using self-organizing maps. Additional modeling research was conducted with the infill density, surface area, and volume ratio identified as key elements contributing to the results. Enhanced dissolution was achieved following thorough testing and ANN modeling in addition to validation(46).

Application of AI for the Detection of Tablet Defects

The use of AI in identifying tablet defects has transformed quality control methods in pharmaceutical production. AI algorithms and computer vision techniques are utilized to examine images of tablets, facilitating the automated and effective detection of defects such as cracks, chips, discoloration, or differences in shape and size. By training AI models on extensive datasets of labeled images, the system acquires the ability to accurately categorize and recognize various types of defects, achieving high levels of precision and recall(47). Traditional techniques, such as X-ray computed tomography, have been employed to examine the internal structure of tablets, yet they remain time-consuming and hinder the demand for the swift production of tablets. Deep learning is applied alongside X-ray tomography to identify tablet defects. Ma et al. investigated the use of neural networks for tablet defect detection aided by image analysis performed through X-ray tomography. These investigators have created multiple batches of tablets by incorporating excipients like microcrystalline cellulose along with mannitol(48). The prepared batches were evaluated using what is known as the image augmentation technique. Three distinct models were employed during the same study, including UNetA, which is useful for differentiating the characteristics of tablets from those of bottles. Module 2 was utilized for the recognition of individual tablets through augmented analysis. The internal fissures within the tablet's internal structure were examined using UNetB. Such UNet networks have been utilized to assess tablet defects with greater precision, thereby facilitating the detection of defects with notable reductions in time, costs, and workload. This AI-driven detection not only enhances the speed and precision of defect recognition but also diminishes reliance on manual inspection, reducing human errors and subjective evaluation. The real-time observation features of AI systems guarantee the quick identification of flaws, enabling swift action and stopping the distribution of defective tablets in the market. In the end, the incorporation of AI in tablet flaw detection improves product quality, boosts productivity, and guarantees the safety and effectiveness of pharmaceutical products(49).

AI for the Prediction of Physicochemical Stability

AI has surfaced as a robust instrument for forecasting the physicochemical stability of oral dosage forms in pharmaceutical studies. By utilizing machine learning algorithms and computational models, AI is capable of examining and interpreting extensive datasets, encompassing drug attributes, formulation specifics, and ecological conditions, to anticipate the stability of oral formulations. AI models can evaluate elements such as drug deterioration, interactions with excipients, and environmental impacts on formulation stability. These forecasting abilities allow researchers to refine formulation designs, detect possible stability concerns early in the development stage, and make knowledgeable choices to improve the shelf life and effectiveness of oral dosage forms. The incorporation of AI into stability forecasting enhances the efficiency and economic viability of drug development procedures, ultimately resulting in the provision of safe and effective medications for patients(50). Certain researchers have examined the use of machine learning for assessing solid dispersion utilizing various algorithms. Han et al. investigated the use of machine learning for forecasting solid dispersion by deploying ANN in conjunction with K-nearest neighbor (KNN) algorithms and a light gradient boosting machine (LightGBM). The SVM was similarly employed in this context. KNN is a nonparametric form of supervised learning classifier. It was utilized to categorize or finalize the predictions for the grouping along with the individual data point. The free and open-source distributed gradient boosting framework created with machine learning is LightGBM. It is commonly used for ranking tasks and classification along with machine learning applications. In this research, around fifty drug molecules with six hundred forty-six data points related to physical stability were gathered from the public database and utilized for the training model. The creation of the database was done using molecular representations and molecular descriptors, including molecular weight, in addition to the hydrogen bond acceptor count. The melting point and the total number of heavy atoms also served as molecular descriptors. Over a period of three months, an accelerated stability study was carried out to further assess the model's performance as part of the physical stability prediction. They achieved an overall accuracy of 82% for the same experiments(51).

Contribution of AI to Dissolution Rate Predictions

The rate at which a drug dissolves in a biological fluid, known as the dissolution rate, is an essential factor that influences its bioavailability and therapeutic efficacy. AI has played a vital role in predicting dissolution rates, assisting in the improvement of drug formulations and dosage forms. By analyzing large datasets of experimental information, AI models can uncover important physicochemical characteristics and molecular attributes that affect the dissolution process. These models utilize machine learning algorithms to understand intricate patterns and connections between drug properties and dissolution rates, allowing for precise predictions. By offering insights into the dissolution behavior of various drug formulations, AI supports the development of more effective drug delivery systems and aids in choosing optimal formulation strategies to boost drug solubility and absorption. Some research has indicated that solid dispersions remain stable without precipitation when excipients are added(52). Dong et al. investigated a technique for forecasting dissolution and the dissolution rate by utilizing AI for a minimum of 50 active pharmaceutical ingredients in combination with 25 polymers. Several AI algorithms they employed are SVM, LightGBM, and extreme gradient boosting (XGBoost). XGBoost is a scalable library associated with machine learning, which features a distributed gradient-boosted decision tree that aids in predicting issues related to unstructured data, such as images and text. The artificial neural network was implemented to surpass all other varieties of algorithms or frameworks. In the same research, the same group employed molecular computational software to create descriptors for both the active pharmaceutical ingredients and the polymers. The variables chosen for the same study included temperature, drug loading, and volume, whereas dissolution was identified as the binary output consisting of precipitation or supersaturation. The dissolution rate was deemed the research output for the same, leading to improved accuracy in predicting the results for the dissolution profiles of the chosen active pharmaceutical ingredients, as well as the polymers(53).

Limitations of AI

Regardless of the benefits provided by AI-driven models, various limitations warrant thorough assessment. These obstacles, such as the necessity for large datasets, possible biases, and issues related to interpretability, underscore the significance of combining AI with conventional experimental techniques to guarantee drug safety and effectiveness. Here are the key limitations:

• Limited availability of data: AI models significantly depend on large datasets for precise predictions. Nevertheless, restrictions in data accessibility, particularly for uncommon illnesses or particular groups, may cause predictions to be less accurate or biased. Problems such as insufficient demographic representation in training datasets can produce biased outcomes. The quality and representativeness of the data utilized for model development need to be thoroughly evaluated.
• Biases in data: The efficacy of AI models depends on impartial and comprehensive training data. Should biases or gaps be present in the data, predictions could be distorted. In clinical trials, uniformity in patient populations can lead to issues, undermining the model's capacity to accurately forecast drug effectiveness. It is essential to provide representative, reliable, and unbiased data for the advancement of AI models.
• Inability to incorporate new data: Once trained, refreshing AI models with new information can be difficult, especially in drug development where the continual appearance of fresh data is typical. Modifying models demands considerable time and resources, and neglecting to maintain models can lead to incorrect predictions and poor decision-making. Plans for model updates must be thoughtfully constructed.
• Lack of transparency: AI models, frequently seen as "black boxes" because of their intricate algorithms, do not provide clarity in detailing how predictions are made. This lack of clarity presents obstacles in securing regulatory approval, as proving the precision and dependability of the model's predictions becomes challenging. Moreover, the absence of transparency may result in distrust concerning predictions, especially when they do not align with the anticipations of clinicians or researchers.
• Constrained capability in addressing Variability: AI models, educated on large datasets, may face difficulties in accurately predicting drug responses for individuals who differ greatly from the average. This concern is especially important for medications that show varying effects in different patients, where significant variability plays a critical role.
• Interpretation of results: The intricacy of AI models may result in outputs that are difficult to interpret, even for specialists. The absence of clear justifications for predictions creates obstacles for clinicians and researchers in comprehending and converting results into actionable insights. Enhanced interpretability and explainability of AI models are crucial for effective application.
• Ethical considerations: Ethical issues, including patient confidentiality, data ownership, and consent matters, emerge when applying AI in drug development. The gathering and utilization of sensitive health information require meticulous attention to safeguard privacy and honour rights. Regulatory bodies are essential in formulating protocols and guidelines to tackle these ethical factors.
• Lack of clinical expertise: While AI is capable of recognizing correlations, it might find it challenging to grasp the complexities and significant impacts of particular parameters, especially in tailored patient therapies. The statistical characteristics of AI algorithms could restrict their comprehension of intricate factors that affect treatment choices, particularly in situations where personalized elements are crucial(54-57).

CONCLUSION

The incorporation of Artificial Intelligence (AI) is set to bring about a groundbreaking period for the pharmaceutical sector, especially in the areas of drug discovery, development, and delivery. AI-driven virtual screening methods will hasten the examination of extensive chemical libraries, speeding up the discovery of therapeutic candidates and lead compounds. The potential of AI also includes facilitating precision medicine, classifying patients, forecasting therapy responses, and tailoring medications based on genomic, proteomic, and clinical data evaluation. In the development of drugs, AI offers the potential to generate innovative compounds with specific target-binding properties using deep learning and generative models, thus improving the effectiveness of medications and minimizing adverse reactions. Areas where AI algorithms are expected to be crucial include creating dose formulations tailored to individual patients and conducting safety assessments, enhancing medicine formulations, and improving treatment results while forecasting possible side effects and toxicity. Monitoring systems powered by AI, along with wearable devices and sensors, will enable remote care for patients, promoting better adherence to medication through tailored therapy suggestions. The impact of AI also reaches clinical trial design, the selection of patients, and recruitment processes, employing electronic health records and genetic data to find suitable patients, reduce the costs of trials, and accelerate the approval processes. Moreover, the real-time monitoring and control abilities of AI are anticipated to enhance continuous manufacturing processes in the pharmaceutical industry, guaranteeing consistency and effectiveness. The capacity of AI in regulatory decisions is considerable, as it evaluates extensive datasets to guide choices, hastening the approval of medications and improving safety protocols. To summarise, AI is set to revolutionize medication delivery technology by offering targeted and personalized treatments. AI-driven models for pharmacokinetics and pharmacodynamics provide advantages over conventional methods, potentially removing the necessity for animal experiments and clinical trials. The integration of AI technology presents substantial potential for accelerating drug development, enhancing patient results, and propelling the pharmaceutical industry into the next evolutionary phase, from 4.0 to 5.0. While challenges such as data quality, regulatory systems, and ethical concerns need to be tackled, ongoing advancements and collaborations present considerable potential for a transformative shift in pharmaceutical practices.

REFERENCES

1. Bhatt P, Singh S, Kumar V, Nagarajan K, Mishra SK, Dixit PK, et al. Artificial intelligence in pharmaceutical industry: Revolutionizing drug development and delivery. Current Artificial Intelligence. 2024;2(1):E051223224198.
2. Vora LK, Gholap AD, Jetha K, Thakur RRS, Solanki HK, Chavda VP. Artificial intelligence in pharmaceutical technology and drug delivery design. Pharmaceutics. 2023;15(7):1916.
3. Selvaraj C, Chandra I, Singh SK. Artificial intelligence and machine learning approaches for drug design: challenges and opportunities for the pharmaceutical industries. Molecular diversity. 2021:1-21.
4. Sahu A, Mishra J, Kushwaha N. Artificial intelligence (AI) in drugs and pharmaceuticals. Combinatorial chemistry & high throughput screening. 2022;25(11):1818-37.
5. Mak K-K, Pichika MR. Artificial intelligence in drug development: present status and future prospects. Drug discovery today. 2019;24(3):773-80.
6. Patel SS, Shah SA. Artificial intelligence: Comprehensive overview and its pharma application. Asian Journal of Pharmacy and Technology. 2022;12(4):337-48.
7. Sampene AK, Nyirenda F. Evaluating the effect of artificial intelligence on pharmaceutical product and drug discovery in China. Future Journal of Pharmaceutical Sciences. 2024;10(1):58.
8. Blanco-Gonzalez A, Cabezon A, Seco-Gonzalez A, Conde-Torres D, Antelo-Riveiro P, Pineiro A, et al. The role of AI in drug discovery: challenges, opportunities, and strategies. Pharmaceuticals. 2023;16(6):891.
9. Lamberti MJ, Wilkinson M, Donzanti BA, Wohlhieter GE, Parikh S, Wilkins RG, et al. A study on the application and use of artificial intelligence to support drug development. Clinical therapeutics. 2019;41(8):1414-26.
10. Zhavoronkov A. Artificial intelligence for drug discovery, biomarker development, and generation of novel chemistry. ACS Publications; 2018. p. 4311-3.
11. Koromina M, Pandi M-T, Patrinos GP. Rethinking drug repositioning and development with artificial intelligence, machine learning, and omics. Omics: a journal of integrative biology. 2019;23(11):539-48.
12. Noorain, Srivastava V, Parveen B, Parveen R. Artificial intelligence in drug formulation and development: applications and future prospects. Current drug metabolism. 2023;24(9):622-34.
13. Mehta A, Niaz M, Adetoro A, Nwagwu U. Advancements in Manufacturing Technology for the Biotechnology Industry: The Role of Artificial Intelligence and Emerging Trends. International Journal of Chemistry, Mathematics and Physics. 2024;8(2):12-8.
14. Khanna A, El Barachi M, Jain S, Kumar M, Nayyar A. Artificial Intelligence and machine learning in drug design and development: John Wiley & Sons; 2024.
15. Chowdhury LAM, Rana T, Azim MI. Intellectual capital efficiency and organisational performance: In the context of the pharmaceutical industry in Bangladesh. Journal of Intellectual Capital. 2019;20(6):784-806.
16. Ahmed I, Jeon G, Piccialli F. From artificial intelligence to explainable artificial intelligence in industry 4.0: a survey on what, how, and where. IEEE Transactions on Industrial Informatics. 2022;18(8):5031-42.
17. Jiménez-Luna J, Grisoni F, Weskamp N, Schneider G. Artificial intelligence in drug discovery: recent advances and future perspectives. Expert opinion on drug discovery. 2021;16(9):949-59.
18. Jan Z, Ahamed F, Mayer W, Patel N, Grossmann G, Stumptner M, et al. Artificial intelligence for industry 4.0: Systematic review of applications, challenges, and opportunities. Expert Systems with Applications. 2023;216:119456.
19. Shaheen MY. Applications of Artificial Intelligence (AI) in healthcare: A review. ScienceOpen Preprints. 2021.
20. Yang X, Wang Y, Byrne R, Schneider G, Yang S. Concepts of artificial intelligence for computer-assisted drug discovery. Chemical reviews. 2019;119(18):10520-94.
21. 21.      Cho O-h. Analysis of the Impact of Artificial Intelligence Applications on the Development of Accounting Industry. Nanotechnology Perceptions. 2024:74-83.
22. Gupta R, Srivastava D, Sahu M, Tiwari S, Ambasta RK, Kumar P. Artificial intelligence to deep learning: machine intelligence approach for drug discovery. Molecular diversity. 2021;25:1315-60.
23. Greenberg ZF, Graim KS, He M. Towards artificial intelligence-enabled extracellular vesicle precision drug delivery. Advanced Drug Delivery Reviews. 2023:114974.
24. Villaseñor-Cavazos FJ, Torres-Valladares D, Lozano O. Modeling and optimization of nanovector drug delivery systems: exploring the most efficient algorithms. Journal of Nanoparticle Research. 2022;24(6):119.
25. Hassanzadeh P, Atyabi F, Dinarvand R. The significance of artificial intelligence in drug delivery system design. Advanced drug delivery reviews. 2019;151:169-90.
26. He S, Leanse LG, Feng Y. Artificial intelligence and machine learning assisted drug delivery for effective treatment of infectious diseases. Advanced drug delivery reviews. 2021;178:113922.
27. Serrano DR, Luciano FC, Anaya BJ, Ongoren B, Kara A, Molina G, et al. Artificial intelligence (AI) applications in drug discovery and drug delivery: Revolutionizing personalized medicine. Pharmaceutics. 2024;16(10):1328.
28. Wilson B, Km G. Artificial intelligence and related technologies enabled nanomedicine for advanced cancer treatment. Taylor & Francis; 2020. p. 433-5.
29. Habeeb M, You HW, Umapathi M, Ravikumar KK, Mishra S. Strategies of Artificial intelligence tools in the domain of nanomedicine. Journal of Drug Delivery Science and Technology. 2024;91:105157.
30. Aundhia C, Parmar G, Talele C, Shah N, Talele D. Impact of artificial intelligence on drug development and delivery. Current Topics in Medicinal Chemistry. 2024.
31. Alshawwa SZ, Kassem AA, Farid RM, Mostafa SK, Labib GS. Nanocarrier drug delivery systems: characterization, limitations, future perspectives and implementation of artificial intelligence. Pharmaceutics. 2022;14(4):883.
32. Li H, Yu Y, Faraji Dana S, Li B, Lee C-Y, Kang L. Novel engineered systems for oral, mucosal and transdermal drug delivery. Journal of drug targeting. 2013;21(7):611-29.
33. Torres-Suárez AI, Martín-Sabroso C, Fraguas-Sánchez AI, Rojo MÁ, Garrosa M, Fernández-Carballido A. Design of dosage forms: influences of anatomy and administration routes.  Dosage Forms, Formulation Developments and Regulations: Elsevier; 2024. p. 183-221.
34. Garcia-del Rio L, Diaz-Rodriguez P, Landin M. Design of novel orotransmucosal vaccine-delivery platforms using artificial intelligence. European Journal of Pharmaceutics and Biopharmaceutics. 2021;159:36-43.
35. Wu K, Li X, Zhou Z, Zhao Y, Su M, Cheng Z, et al. Predicting pharmacodynamic effects through early drug discovery with artificial intelligence-physiologically based pharmacokinetic (AI-PBPK) modelling. Frontiers in Pharmacology. 2024;15:1330855.
36. Pawar V, Patil A, Tamboli F, Gaikwad D, Mali D, Shinde A. Harnessing the power of AI in pharmacokinetics and pharmacodynamics: A comprehensive review. AAPS PharmSciTech. 2021;14(2):426-39.
37. Mansour RF, El Amraoui A, Nouaouri I, Díaz VG, Gupta D, Kumar S. Artificial intelligence and internet of things enabled disease diagnosis model for smart healthcare systems. IEEE Access. 2021;9:45137-46.
38. Rezaei M, Rahmani E, Khouzani SJ, Rahmannia M, Ghadirzadeh E, Bashghareh P, et al. Role of artificial intelligence in the diagnosis and treatment of diseases. Kindle. 2023;3(1):1-160.
39. Kumar Y, Koul A, Singla R, Ijaz MF. Artificial intelligence in disease diagnosis: a systematic literature review, synthesizing framework and future research agenda. Journal of ambient intelligence and humanized computing. 2023;14(7):8459-86.
40. Shen J, Zhang CJ, Jiang B, Chen J, Song J, Liu Z, et al. Artificial intelligence versus clinicians in disease diagnosis: systematic review. JMIR medical informatics. 2019;7(3):e10010.
41. Kanwal K, Ibrahim Z, Siddiqui AA. The Convergence of Artificial Intelligence, Digital Therapeutics, and Biomedical Electronics in the field of Speech-Language Therapy: AI and Digital Innovations in Speech Therapy. Allied Medical Research Journal. 2024;2(2):277-92.
42. Remus A, Tadeo X, Kai GNS, Blasiak A, Kee T, Vijayakumar S, et al. CURATE. AI COR-Tx platform as a digital therapy and digital diagnostic for cognitive function in patients with brain tumour postradiotherapy treatment: protocol for a prospective mixed-methods feasibility clinical trial. BMJ open. 2023;13(10):e077219.
43. Washington P, Kalantarian H, Kent J, Husic A, Kline A, Leblanc E, et al. Improved digital therapy for developmental pediatrics using domain-specific artificial intelligence: machine learning study. JMIR pediatrics and parenting. 2022;5(2):e26760.
44. Hu J, Wan J, Xi J, Shi W, Qian H. AI-driven design of customized 3D-printed multi-layer capsules with controlled drug release profiles for personalized medicine. International Journal of Pharmaceutics. 2024;656:124114.
45. Jiang J, Ma X, Ouyang D, Williams III RO. Emerging artificial intelligence (AI) technologies used in the development of solid dosage forms. Pharmaceutics. 2022;14(11):2257.
46. Elbadawi M, McCoubrey LE, Gavins FK, Ong JJ, Goyanes A, Gaisford S, et al. Harnessing artificial intelligence for the next generation of 3D printed medicines. Advanced Drug Delivery Reviews. 2021;175:113805.
47. Vijayakumar A, Vairavasundaram S, Koilraj JAS, Rajappa M, Kotecha K, Kulkarni A. Real-time visual intelligence for defect detection in pharmaceutical packaging. Scientific Reports. 2024;14(1):18811.
48. Diószegi A, Ficzere M, Mészáros LA, Péterfi O, Farkas A, Galata DL, et al. Automated tablet defect detection and the prediction of disintegration time and crushing strength with deep learning based on tablet surface images. International Journal of Pharmaceutics. 2024;667:124896.
49. Quan HT, Huy DD, Hoan NT, Duc NT, editors. Deep learning-based automatic detection of defective tablets in pharmaceutical manufacturing. 8th International Conference on the Development of Biomedical Engineering in Vietnam: Proceedings of BME 8, 2020, Vietnam: Healthcare Technology for Smart City in Low-and Middle-Income Countries; 2022: Springer.
50. Points LJ, Taylor JW, Grizou J, Donkers K, Cronin L. Artificial intelligence exploration of unstable protocells leads to predictable properties and discovery of collective behavior. Proceedings of the National Academy of Sciences. 2018;115(5):885-90.
51. Leniak A, Pietrus W, Kurczab R. From NMR to AI: designing a novel chemical representation to enhance machine learning predictions of physicochemical properties. Journal of Chemical Information and Modeling. 2024;64(8):3302-21.
52. Nagy B, Petra D, Galata DL, Démuth B, Borbás E, Marosi G, et al. Application of artificial neural networks for Process Analytical Technology-based dissolution testing. International journal of pharmaceutics. 2019;567:118464.
53. Wang H, Kwong CF, Liu Q, Liu Z, Chen Z. [Retracted] A Novel Artificial Intelligence System in Formulation Dissolution Prediction. Computational Intelligence and Neuroscience. 2022;2022(1):8640115.
54. Chowdhury M, Sadek AW. Advantages and limitations of artificial intelligence. Artificial intelligence applications to critical transportation issues. 2012;6(3):360-75.
55. Hauer T. Importance and limitations of AI ethics in contemporary society. Humanities and Social Sciences Communications. 2022;9(1):1-8.
56. Smoke S. Artificial intelligence in pharmacy: A guide for clinicians. American Journal of Health-System Pharmacy. 2024:zxae051.
57. Wong A, Wentz E, Palisano N, Dirani M, Elsamadisi P, Qashou F, et al. Role of artificial intelligence in pharmacy practice: a narrative review. Journal of the American College of Clinical Pharmacy. 2023;6(11):1237-50.