Hybrid AI Innovation: The pursuit of new pharmaceutical treatments is one of humanity’s most essential and timeless endeavors in improving global health. However, this search is far from easy, often hindered by the complexity of biological systems, the high cost of research, and the rigorous demands of regulatory approval. The road from idea to market is often long, expensive, and fraught with inefficiencies and high failure rates. These challenges have spurred innovation in drug discovery, and over the last decade, Artificial Intelligence (AI) and Machine Learning (ML) have emerged as game-changers in the field.
These technologies are not only speeding up the process but also improving the precision of drug development, offering a new paradigm that could ultimately make the development of essential therapies more efficient, cost-effective, and faster.
Current State of Inefficiency in the Drug Development World
The process of drug discovery and development is complex and time-consuming, despite decades of scientific advancements. Historically, the development of new drugs has taken many years, often a decade or more, from initial discovery to the point where a drug is available for public use. Furthermore, over 90% of drug candidates that enter clinical trials do not make it to the market, a staggering statistic that underscores the challenges of successfully translating laboratory findings into real-world therapies.
A major reason behind this high failure rate is the inherent unpredictability of how a drug will perform in humans. While the preclinical phase—comprising in vitro studies and animal testing—serves as a critical step in the drug development process, these models often fail to accurately replicate the complexities of human biology. Animal models, while useful, have long been criticized for their inability to predict human responses reliably. The biological differences between animals and humans mean that drugs can behave in ways that are difficult to anticipate.
For instance, a drug that shows promise in rodents or primates may fail in humans due to different metabolic processes, immune responses, or other physiological factors that cannot be accurately modeled in animals. As a result, many potential treatments that show promise early on in the discovery process end up failing during the clinical trial phases, leading to wasted time, resources, and significant financial losses.
The inefficiencies in the drug development process also extend to the limitations of current data analysis methods. Traditional approaches to data analysis often rely on small, isolated datasets or manual analysis of trial results, which can be time-consuming and prone to human error. Furthermore, these approaches may not fully leverage the vast amount of data generated during clinical trials, leaving valuable insights untapped. As a result, drug developers may miss opportunities to optimize treatment regimens or identify potential side effects early on in the process. The inability to process large volumes of complex data quickly and accurately is another key inefficiency in the current system.
Finally, traditional drug development has typically focused on creating a “one-size-fits-all” treatment, but this approach often fails to account for the genetic, environmental, and lifestyle differences that influence how individuals respond to medications. As a result, many patients may not respond to treatments as expected, leading to trial and error in finding the right therapy. This lack of precision in drug development not only hinders the effectiveness of treatments but also increases the time and cost of developing drugs that work for a broader population.
The Impact of AI in Drug Discovery in 2025
AI and ML are beginning to take center stage in transforming the way drugs are discovered, tested, and brought to market. One of the most notable signs of this change is the growing number of regulatory submissions involving AI/ML technologies. The U.S. Food and Drug Administration (FDA) has seen a substantial increase in regulatory submissions that incorporate AI and ML in the past decade, indicating a growing confidence in the use of these technologies to support drug development.
These technologies are now being applied in various stages of drug development, from the initial discovery phase and drug repurposing to clinical trial design, dosage optimization, and even post-market surveillance. The diversity of therapeutic areas benefiting from AI/ML applications—ranging from oncology to infectious diseases—illustrates the immense potential of these technologies to address some of the most complex challenges in modern healthcare.
In response to the widespread adoption of AI/ML in drug development, the FDA has been actively working to improve its capacity to evaluate and oversee these technologies. In 2025, the FDA released draft guidance titled Considerations for the Use of Artificial Intelligence to Support Regulatory Decision Making for Drug and Biological Products, which offers industry stakeholders practical recommendations on how to leverage AI to create data or insights that can inform regulatory assessments of drug safety, effectiveness, and quality.
This guidance is a crucial step toward ensuring that AI and ML applications in drug development meet regulatory standards and can be trusted to provide reliable and valid results. This is not the first time the FDA has addressed this topic; another example is a discussion paper from several years back titled “Using Artificial Intelligence & Machine Learning in the Development of Drug and Biological Products,” aimed at fostering a dialogue with the community through comments and feedback.
Parallel to the FDA’s initiatives, the European Medicines Agency (EMA) has also recognized the importance of AI in the pharmaceutical industry. At the end of 2023, the EMA released a reflection paper outlining how AI could be integrated across the medicinal product lifecycle. This document, titled Artificial intelligence workplan to guide use of AI in medicines regulation, highlights the agency’s commitment to understanding and incorporating AI/ML innovations into regulatory frameworks. By developing guidelines for the use of AI in the drug development process, the EMA seeks to ensure that AI technologies are applied in a way that guarantees the safety and efficacy of new medicines.
These regulatory efforts by the FDA and EMA underscore the growing consensus that AI and ML will play an essential role in shaping the future of drug development.
New Technologies Pave the Way for More Available Treatments
Hybrid AI, an advanced technology that combines predictive machine learning algorithms with biologically driven simulation models, is transforming drug discovery and development. This dual approach harnesses the power of AI to analyze vast datasets and predict drug interactions while leveraging simulations to model complex biological processes in a controlled, virtual environment. As a result, Hybrid AI offers more accurate, mechanistically explainable insights that neither traditional drug discovery methods nor AI alone can achieve.
For example, in the preclinical phase, researchers can simulate drug interactions within the human body while predicting how drug candidates behave across biological systems. By modeling pharmacokinetics (PK)—how a drug is absorbed, distributed, metabolized, and excreted—and pharmacodynamics (PD)—how the drug affects the body—Hybrid AI enables a deeper understanding of potential drug efficacy and safety before clinical trials even begin. This early-stage risk mitigation allows researchers to filter out suboptimal candidates faster, accelerating the drug discovery process while reducing costly failures in later stages.
Beyond predicting efficacy, Hybrid AI also identifies potential side effects and toxicities by analyzing metabolite interactions. Metabolites, or byproducts formed as the body processes a drug, can sometimes contribute to unexpected toxicity, particularly if they accumulate in certain organs. By combining predictive models of metabolite formation with simulations of organ-specific drug exposure, Hybrid AI provides a more complete risk profile of drug candidates—flagging toxicity risks that traditional methods might overlook.
By using hybrid AI to simulate biological systems and predict drug interactions, researchers can quickly test multiple hypotheses and narrow the focus to the most promising drug candidates. This approach not only reduces the risk of failure in later stages of development but also accelerates the overall drug discovery process, enabling therapies to reach patients more quickly.
One of the most critical advantages of Hybrid AI is its ability to enhance explainability in drug discovery. Unlike AI models, which provide predictions with sub-clear justifications, Hybrid AI integrates mechanistic simulations, offering insights that are grounded in biological processes rather than statistical correlations alone. This reduces bias, increases regulatory confidence, and enables better-informed decision-making, particularly in high-stakes fields like oncology and rare disease treatment.
Conclusion
The integration of hybrid AI into the drug development process represents a transformative shift that will not only accelerate the discovery of new therapies but also enhance their precision and safety. By combining AI’s predictive power with simulation-based modeling, hybrid AI offers new ways to predict the success of drug candidates, identify potential side effects, and optimize treatment regimens. As demonstrated by recent case studies, hybrid AI has the potential to reduce the risk of failure in drug development, shorten development timelines, and reduce costs—ultimately making life-saving therapies available to patients more quickly. This integrated approach to drug discovery is not just a technological advancement; it is a paradigm shift that promises to reshape the future of medicine.
Editor’s note: Dr. Jo Varshney is leading pharmaceutical innovation as the Founder and CEO of VeriSIM Life and the inventor of the groundbreaking BIOiSIM® technology. This technology revolutionizes preclinical drug development by providing the industry’s only AI-powered “credit score” for preclinical assets, significantly accelerating the delivery of cost-efficient, life saving therapies. A respected leader with numerous accolades including San Francisco Business Times’s Most Influential Women in Bay Area Business, Dr. Varshney has spearheaded high-impact collaborations with top-tier pharmaceutical firms, government bodies, and leading academic and medical institutions. She holds a DVM and a Ph.D. in Comparative Oncology/Genomics from the University of Minnesota, and further graduate degrees in Comparative Pathology and Computational Sciences from Penn State and UC San Francisco, respectively.