Murray McKinnon, Ph.D., Chief Scientific Officer, Empress Therapeutics
AI is undeniably transforming drug development, but we’re still in the early stages of realizing its full potential. Recent advances show that AI is moving beyond hype, as tools now analyze vast data sets, identify novel targets, and even model complex biological pathways. These capabilities, highlighted in recent breakthroughs, are enabling faster, more cost-effective drug discovery.
Companies like Empress are pushing the boundaries by using AI to uncover previously unknown biological pathways and drug targets. For example, Empress’s unique Chemilogics approach applies proprietary computational models to map intricate disease mechanisms, enabling the discovery of therapies that target diseases at their root. This could lead to treatments that are not only more effective but also better understood, reducing development risks and increasing the likelihood of clinical success.
While challenges like data quality and integration into existing workflows remain, they are becoming manageable as more interdisciplinary teams collaborate and AI tools become increasingly refined. The regulatory landscape is also evolving to embrace these innovations.
The real value of AI lies in its ability to explore what was once considered undruggable, unlocking new opportunities for innovation. By leveraging its power responsibly, we can accelerate the development of therapies that improve patients’ lives worldwide.
Richard Lee, Ph.D., Director, Core Technology, ACD/Labs
AI is influencing drug development, but its impact is more practical evolution than revolutionary transformation. In early-stage drug discovery, AI tools like DeepMind’s AlphaFold have greatly advanced protein structure prediction, while a growing number of AI-driven startups are emerging. AI has significant potential to enhance efficiency and accelerate timelines by streamlining data analysis and facilitating hypothesis generation, while complementing traditional approaches.
In practice, AI excels in well-defined, data-rich areas, such as virtual screening of compounds, biomarker identification, and optimizing clinical trial design. Its value lies in augmenting decision-making rather than providing standalone solutions. AI can quickly process large data sets, uncovering patterns and insights that would take longer, but its outputs still require validation through established scientific methods.
Challenges, such as inconsistent data quality, integration into established workflows, and regulatory acceptance, temper expectations. Rather than expecting disruptive breakthroughs, the industry is increasingly focusing on pragmatic applications, like improving existing processes and decision-making tools.
The outlook for AI in drug development is promising but should not be viewed as a magic bullet. It is best seen as a powerful enabler that complements human expertise, helping to streamline R&D, manage risks, and ultimately improve patient outcomes.
Alan Marcus, Chief Growth Officer, LabVantage Solutions
AI is undeniably transforming drug development, but its full potential is still unfolding. While there’s plenty of hype, real progress is being made, particularly in how AI-powered SaaS (software as a service) solutions accelerate research, improve collaboration, and optimize decision-making.
Embedded AI agents are helping researchers sift through vast data sets to identify promising drug candidates faster and with greater accuracy. AI-driven insights are also enhancing clinical trial design by predicting patient responses and optimizing trial site selection.
We’re already seeing SaaS with embedded AI making a tangible impact in collaborative R&D. Virtual research organizations (VROs) leverage SaaS platforms to connect researchers from different organizations, allowing them to work together on drug discovery projects in real-time. AI agents within these platforms analyze multi-source data, identify drug targets, and even assist in drug design. SaaS-based data-sharing platforms are also enabling pharma companies, biotech firms, and academic institutions to pool their data for collective insights — accelerating discoveries that might otherwise take years.
However, challenges remain. Data privacy and security are critical, especially in collaborative environments where multiple organizations share sensitive information. And it’s important to understand that the quality of AI predictions depends on the integrity of the data used to train models. Ethical considerations — such as algorithmic bias and responsible AI use in healthcare — must also be addressed.
Christian Olsen, Associate Vice President, Industry Principal, Biologics, Dotmatics
AI is steadily transforming drug development, with companies like Dotmatics leading the way, providing scientific informatics solutions, integrating data, workflows, and collaboration to accelerate drug discovery. AI and machine learning (ML) are enhancing drug target identification, molecule design, and data analysis. This year, breakthroughs in biotech will focus on digital transformation, data harmonization, and AI/ML, particularly in high-throughput applications for R&D, development, and manufacturing.
Key challenges remain in technological integration, precision medicine, and AI/ML frameworks. Biotech companies must continue investing in AI and ML to stay competitive, as seen with examples like AlphaFold3 in protein prediction. The surge in AI and digital health tech funding signals the importance of these tools in biotech operations.
Precision medicine, which tailors treatments based on genetics, lifestyle, and environment, will become central to healthcare. Electronic health records (EHRs) will advance to become "genome-aware," making pharmacogenomics more accessible. This will enable more precise matching of patients with effective therapies.
AI/ML is also reshaping clinical research, optimizing drug screening, and improving clinical trials by processing large data volumes with greater speed and accuracy. The future will see greater integration of AI/ML models in drug development, from preclinical screening to clinical outcome analysis, advancing targeted therapies and improving patient outcomes.
Mark Kiel, M.D., Ph.D., Co-founder and Chief Scientific Officer, Genomenon
AI is undeniably transforming drug development, but its impact can be viewed in two key ways. Is AI simply accelerating existing processes, or is it enabling entirely new possibilities? In reality, by making traditional workflows faster and more efficient, AI is also unlocking innovations that were previously out of reach.
AI’s ability to handle vast amounts of biomedical data — organizing, annotating, and interpreting information with unprecedented speed — has already revolutionized many aspects of drug discovery.
While there is considerable hype around AI, there is also real measurable value and it lies in its ability to tackle tasks once considered time- and cost-prohibitive, such as analyzing real-world evidence (RWE) data across multiple health systems or computationally mapping complex 3D protein structures. The shift is happening now — moving from bold expectations to concrete results, where real-world advancements, validated outcomes, and practical applications define its success.
Challenges remain, particularly in regulatory adaptation, data accessibility, and system interoperability. Yet, breakthroughs like AI-driven protein modeling, biomarker identification, optimized drug candidate selection, and large-scale genomic analysis continue to push the boundaries of what is possible.
AI won’t replace existing drug development methods, but it will redefine them — making processes faster, smarter, and more precise than ever before.
Simon Wagschal, Ph.D., Associate Director, Advanced Chemistry Technologies, Lonza Small Molecules
As is true for all industries today, AI is driving the drug development and manufacturing industry forward, though it's important to differentiate between its substantial potential and the current reality.
AI has already made a tremendous impact in revolutionizing drug development. One major breakthrough is leveraging AI to help navigate the increasing complexity of drug candidates, particularly small molecule APIs. For example, the growing number of synthetic steps required to produce these drugs creates challenges in lead times and raw material management. Technological solutions, like Lonza’s AI-enabled Route Scouting Service, provide synthetic pathways that are more supply chain resilient and offer insights for optimal route design, accelerating API synthesis for both clinical and commercial manufacturing.
However, significant barriers remain including risk of inaccurate models, data privacy and infringement, and regulatory hurdles. It’s essential for drug developers and their partners to ensure availability of quality data for more comprehensive models and appropriate guardrails are in place for data protection and collection.
While AI is still in the infancy stage, the transformation is underway and there are numerous avenues for opportunity in the future. To fully realize the potential of AI in drug development, continued investment in data infrastructure, development of robust AI models, and collaborative efforts is needed.
Jo Varshney, Ph.D., Chief Executive Officer and Founder, VeriSIM Life
AI is making significant strides in drug development, but the industry is still navigating the gap between promise and practical impact. While many AI-driven approaches focus on identifying potential drug candidates through chemical structure analysis and target binding predictions, only a few are taking the next step — incorporating biological complexity into AI models to improve translation to clinical success.
The challenge lies in the fact that drug efficacy and safety are not solely determined by molecular interactions at the target level. A drug’s success depends on how it behaves in a living system — how it is absorbed, distributed, metabolized, and excreted, and how it interacts with different tissues and genetic factors. Traditional AI models, often trained on historical data sets, struggle to capture these dynamic biological processes, leading to high failure rates when drugs move from preclinical to clinical stages.
Innovative companies are moving beyond chemistry-first AI by integrating systems biology, pharmacology, and advanced simulation techniques to model how drugs interact with the body before they enter human trials. These approaches aim to reduce late-stage failures by predicting toxicity, metabolism, and patient-specific responses earlier in the pipeline.
While AI holds immense potential in drug development, its transformative impact will depend on how well models account for the biological and physiological realities of human disease. Without this, AI risks becoming another tool that optimizes existing processes rather than fundamentally improving the predictability of drug success. The next phase of AI-driven drug development will likely be defined by how well companies bridge this gap between computational predictions and real-world biology.
One of the biggest breakthroughs will continue to be the integration of multi-scale biological data into AI models. Moving beyond traditional chemistry-based predictions, AI is now incorporating genomics, proteomics, and metabolomics to improve accuracy in predicting drug efficacy and safety. Additionally, predictive digital twins — computational models that simulate human physiology and disease progression — are enabling researchers to test drug responses virtually, reducing reliance on animal models and improving early-stage decision-making.
Another major advancement will be AI-driven drug repurposing or redirecting, which identifies new therapeutic uses for existing drugs by analyzing vast clinical and pharmacological data sets. This approach can significantly reduce development costs and timelines. AI is also playing a critical role in personalized and precision medicine, helping to stratify patients based on biomarkers and genetic data. By matching drugs to the right patient populations, AI has the potential to improve clinical trial success rates. Furthermore, AI-driven toxicity and drug exposure predictions are helping to eliminate non-viable candidates early, reducing costly late-stage failures.
Despite these advancements, significant barriers remain. Many AI models still rely on historical data sets and chemical structure-based predictions, which do not fully capture the complexity of human biology. The quality, availability, and diversity of biological data also pose challenges, as incomplete or biased datasets can limit model accuracy. Additionally, regulatory uncertainty remains a major hurdle, with agencies still defining how AI-generated predictions should be validated and incorporated into the approval process.
Another challenge is the high computational demands of simulating complex biological systems. AI models require large-scale computing power and extensive validation, increasing costs and slowing adoption. Lastly, the pharmaceutical industry's resistance to change presents a cultural barrier. Traditional R&D methods remain deeply embedded, and integrating AI-driven insights into existing workflows requires trust, validation, and alignment with industry standards.
Looking ahead, AI's impact on drug development will depend on how well it integrates biological complexity, overcomes data quality limitations, and gains regulatory and industry acceptance. The most transformative approaches will be those that combine computational modeling with experimental validation, ensuring that AI-driven predictions translate into real-world clinical success.
Seongil Cho, Vice President of IT, Samsung Biologics
In my view, the data collection process itself, let alone AI tools that are trained on the data, is still in its development stage. There is room for improvement in the method by which data are collected and the history that the data have undergone to be made more clear. When such seamless availability of bioprocess data becomes universal, the data would shed light not only on understanding the common aspects shared across monoclonal antibodies but also on specific biologic products.
The machine learning expertise of the chemical industry’s diverse manufacturing processes offers many opportunities for the life sciences industry. These opportunities include taking preemptive measures to maintain equipment and analyzing the root cause of quality issues that may arise.
AI holds great potential for the biopharmaceutical and biotechnology industry. Life sciences have taken longer to reach the stages of automation and systematization compared with other industries due to the relatively greater effort required for collecting data, in addition to regulatory safeguards to ensure patient safety. To enhance comprehension of bioprocess characteristics, there needs to be a journey to link the fundamental knowledge of proteins with the outcome of biologic development and manufacturing. Considering the specialized nature of drug development and manufacturing data, new ventures and long-term endeavors are key.
John Lee, Ph.D., Global Head of Cell & Gene Therapy, SK pharmteco
AI is definitely reshaping drug development, but we should remember where it sits on the Gartner Hype Cycle. Right now, it’s likely perched between the Peak of Inflated Expectations — where AI seems like a cure-all — and the Trough of Disillusionment, where its limitations become clear. On the upside, machine learning and computational modeling help researchers analyze massive data sets — like patient genomics — to accelerate target identification and drug design. Generative AI also shows promise for creating novel compounds faster than traditional methods.
However, regulatory questions remain significant. Agencies such as the FDA are working out how to regulate AI-based solutions, especially “black-box” algorithms with limited transparency. This uncertainty can stall adoption, even if the underlying technology is sound. Data quality is another challenge, because AI depends on clean, comprehensive data sets. And for smaller players, implementing advanced platforms can be cost prohibitive.
All things considered, the stage is set for AI to move beyond the Trough of Disillusionment and into the Slope of Enlightenment, where it becomes more transparent and broadly adopted. While that may still be a few years away, the potential for AI to transform drug development remains very real.
Julie Frearson, Ph.D., Senior Vice President and Chief Scientific Officer, Charles River Laboratories
From a drug discovery viewpoint, AI has become an umbrella term for everything predictive and generative and is being integrated into discovery processes across many organizations. The initial hyperbolic claims we saw about how AI would change drug discovery have subsided and instead been replaced with a distinct call for results: is AI discovering drugs? Well, AI-based tools are becoming commonplace in small molecule design and have definitely changed the landscape for protein structure prediction and engineering; however, the impact on modelling complex biological processes to identify new target–disease relationships or predict toxicities remains more theoretical. Overall, we expect AI to have a positive impact on drug discovery, shortening timelines for best in class programs and perhaps increasing the optionality around lead series and development candidates, but the computational algorithms deployed will become increasingly democratized, and the winners will be those organizations with large, high-quality, intentionally developed data sets directed at a suite of models acting in co-ordination with human drug discovery expertise.
Venu Mallarapu, Vice President of Global Strategy & Operations, eClinical Solutions
Companies across sectors are rethinking operations to adapt to an AI-driven landscape. However, biopharma’s AI adoption has been slow, with many still in exploratory stages. Skepticism persists following initial industry hype, and technology alone isn’t a fix – change management and transformed processes are essential for ROI. Organizations also face concerns like regulatory compliance and data privacy.
Despite these challenges, there is sustained interest in AI’s potential to automate and reduce cycle times within R&D. Generative AI presents quick wins for the industry. Tools like chatbot assistants have shown significant success in these areas, accelerating drug discovery, optimizing clinical trials, and enhancing data analysis.
However, many companies face barriers after the proof-of-concept (PoC) stage, due to lack of data, investment, and hard-to-prove outcomes. Mistrust of technology vendors, regulatory issues, and the need for robust infrastructure also remain hurdles.
Modern data infrastructure and analytics enable faster insights and better decision-making, and embedding AI streamlines these processes further. The FDA’s January 2025 draft guidance on AI use is a key milestone for responsible AI adoption. While balancing innovation with caution is crucial, there is a unique opportunity to embed AI across the life cycle, reducing cycle times and delivering therapies to patients faster.
Ann Beliën, Ph.D., Founder and Chief Executive Officer, Rejuvenate Biomed
AI is undeniably transforming drug development and therapeutic innovation, from target identification and drug discovery acceleration to informing personalized medicine.
Thanks to AI, researchers can navigate a vast search space of complex biological interactions, gaining a much deeper understanding of how compounds interact at the molecular level and revealing hidden patterns and biological networks that traditional algorithmic solutions might miss. This is key to identifying treatment opportunities for multifactorial conditions like age-related diseases, which require a multi-target approach through combination products. This added layer of complexity demands novel AI solutions.
In our case, AI is a critical driver of our drug development process, helping identify the most effective and safe drug combinations, significantly reducing development risk and time, and enhancing clinical success rates to bring better therapies to patients faster.
Key challenges in using AI for drug development include the creation and identification of robust, high-quality data, the continuous validation of AI algorithms to prevent bias, and the development of (cost-) appropriate computational environments and insight-generating tools (explainability of outcome). Additionally, supporting regulatory bodies in establishing standardized guidelines and frameworks that support drug approval will be crucial.