Developing new drugs has traditionally been challenging, expensive, and time-consuming. It often relies on the expertise of drug developers and involves a great deal of trial and error. But now, with the emergence of AI technologies, particularly advanced language models and generative AI, this landscape is beginning to change.

The integration of AI into drug discovery is transforming the pharmaceutical world. These cutting-edge tools are opening exceptional opportunities—like speeding up the process of developing new medications, cutting down on costs, and boosting the chances of success for those new drugs hitting the market. Nevertheless, this journey also faces challenges that need to be addressed to fully understand their potential.

This article discusses the challenges, opportunities, strategies, and prospects of AI in drug discovery.

Challenges

Complexity of biological systems

One of the top challenges in drug discovery is the intrinsic complexity of biological systems. Living organisms are complicated  webs of connections at the molecular, cellular, and organ levels. Modeling this elaborate networking accurately is crucial for predicting how a drug will behave in the body. AI systems encounter difficulties in understanding multifaceted and non-linear relationships in biological data¹. Additionally, the dynamic nature of biological structures, which can change their behavior in response to different stimuli, adds to this challenge. As a result, it is complex to create static models that genuinely reflect real-world conditions.

Data quality and availability

Data availability and data quality are also considerable obstacles. The effectiveness of AI in drug discovery is highly dependent on data quality and completeness. Pharmaceutical datasets are frequently tainted by errors, bias, or missing values, which can considerably reduce the predictive accuracy of AI models. Integrating heterogeneous data from numerous sources into a unified platform is a primary challenge. On top of that, the proprietary nature of pharmaceutical data adds another layer of complication. This limits how much access researchers have to the comprehensive datasets that are crucial for training AI and machine learning(ML) models².

Regulatory and ethical concerns

The inclusion of AI into drug discovery raises important regulatory and ethical questions. Ensuring that AI-driven choices are transparent and understandable is essential for getting regulatory approval and trust from the general public³. The “black-box” nature of most AI algorithms can obstruct the ability to know how and why decisions are made, which is a major barrier to regulatory acceptance⁴. In addition, the use of patient data must comply with strict data protection guidelines, and there are ethical considerations around the potential for AI to perpetuate existing biases in healthcare.

Interdisciplinary collaboration

Successful implementation of AI in drug discovery requires collaboration between multiple specialists, including biologists, chemists, data scientists, and AI experts⁵. Bridging the gap between these disciplines can be challenging but is vital for developing robust AI models.

Opportunities

Accelerating drug development

Despite these obstacles, AI and ML offer immense possibilities in this area. One of the most significant advantages is the potential to accelerate drug development. AI and ML can rapidly scan extensive databases of chemical compounds, biological interactions, and disease pathways, cutting the time and cost of bringing new drugs to the market⁶. This can be incredibly helpful in responding to new public health threats, like pandemics, where speed is paramount.

Improving success rate

AI can also improve the success rate of clinical trials by identifying drug candidates more precisely. Traditional drug discovery procedures rely heavily on trial and error, with most candidates failing in late-stage clinical trials. Moreover, AI can forecast side effects and optimize drug formulation, increasing the rate of success⁷. This can lead to better resource allocation and higher return on investment for pharmaceutical companies.

Personalized medicine

One more area where AI can significantly impact is personalized medicine. AI can identify the optimal treatment for specific individuals based on patient information, resulting in targeted therapies and enhanced patient outcomes. It can also locate patient subgroups that are more likely to react to particular medicines, providing more efficient and precise care⁸.

Drug repurposing

Furthermore, drug repurposing is a valid alternative that AI could facilitate. AI can detect new therapeutic applications for these compounds by analyzing existing drugs and their modes of action. This approach can save time and money compared to developing entirely new drugs from scratch, as existing medications have already undergone extensive safety testing⁹. AI-driven drug repurposing has the potential to bring new treatments to market more quickly and at a lower cost, particularly for severe, widespread conditions with limited effective therapies, including Alzheimer’s disease.

Strategies

Different tactics can be adopted to overcome the limitations of AI and ML in drug discovery. Data integration and augmentation are necessary to enhance the robustness of AI models. Pharmaceutical companies could invest in methods to enhance data quality and integrate several datasets into a unified framework¹⁰. This can bypass the flaws of incomplete or biased data and improve the predictive accuracy of AI algorithms.

Explainable AI (XAI) is a research area that aims to make AI algorithms more understandable by clarifying the reasoning behind their decisions and predictions. In this context, developing XAI models is imperative for regulatory approval and ensuring transparency in AI-driven decisions¹¹. This field can assist stakeholders in understanding the rationale behind AI predictions and building trust. This is especially important in the pharma industry, where regulatory bodies demand a clear understanding of how decisions are made.

Combining AI with standard experimental methods can boost the drug discovery process. AI can generate hypotheses that can be validated through lab experiments, establishing an interactive system that employs the strengths of both AI and traditional approaches. This combination can lead to more potent and robust drug development¹².

Promoting interdisciplinary training and collaboration helps connect various fields involved in drug discovery. By fostering a collaborative culture, pharmaceutical companies can leverage the knowledge of experts to develop more effective AI models and innovative solutions.

The future of AI and ML in drug discovery

Modern AI techniques like deep learning and reinforcement learning are promising for drug discovery and development in the coming years. These systems can also improve the accuracy and efficiency of these processes, facilitating the discovery of new drug candidates and therapy targets. The concept of digital twins, where virtual models of biological systems are created, is also very encouraging¹³. Digital twins can simulate drug effects in virtual patients, allowing scientists to predict outcomes more accurately and reduce animal and human testing.

Furthermore, integrating AI with genomics and proteomics can result in a deeper understanding of disease mechanisms and identifying novel drug targets¹⁴. This may create opportunities for designing more effective drugs and personalized medicine approaches. As AI continues to evolve, addressing ethical concerns will be key. Ensuring AI systems are fair, transparent, and accountable will be critical to their universal use in drug discovery.

The future of AI in drug discovery will be marked by collaborative environments where pharmaceutical companies, academia, and tech providers work together. These partnerships can drive innovation and accelerate the delivery of new treatments. By embracing strategic approaches and encouraging cross-functional collaboration, the pharmaceutical industry can unlock the full potential of AI and ML and lead to faster, more efficient, and more targeted drug discovery pipelines.

References

1.    Blanco-González, A.; Cabezón, A.; Seco-González, A.; et al. The Role of AI in Drug Discovery: Challenges, Opportunities, and Strategies. Pharmaceuticals (Basel), 18;16(6):891, 2023.

2.    Zhu, H. Big Data and Artificial Intelligence Modeling for Drug Discovery. Annu Rev Pharmacol Toxicol, 6;60:573-589, 2020.

3.    Mennella, C.; Maniscalco, U.; De Pietro, G.; Esposito, M. Ethical and regulatory challenges of AI technologies in healthcare: A narrative review. Heliyon, 15;10(4):e26297, 2024.

4.    Yang, G.; Ye, Q.; Xia, J. Unbox the black-box for the medical explainable AI via multi-modal and multi-centre data fusion: A mini-review, two showcases and beyond. Inf Fusion, 77:29-52, 2022.

5.    Mirakhori, F.; Niazi, S.K. Harnessing the AI/ML in Drug and Biological Products Discovery and Development: The Regulatory Perspective. Pharmaceuticals (Basel), 3;18(1):47, 2025.

6.    Vora, L.K.; Gholap, A.D.; Jetha, K.; et al. Artificial Intelligence in Pharmaceutical Technology and Drug Delivery Design. Pharmaceutics, 10;15(7):1916, 2023.

7.    Chopra, H.; Annu Shin, D.K.; Munjal, K.; et al. Revolutionizing clinical trials: the role of AI in accelerating medical breakthroughs. Int J Surg, 1;109(12):4211-4220, 2023.

8.    Schork, N.J. Artificial Intelligence and Personalized Medicine. Cancer Treat Res, 178:265-283, 2019.

9.    Yan, C.; Grabowska, M.E.; Dickson, A.L.  et al.  Leveraging generative AI to prioritize drug repurposing candidates for Alzheimer’s disease with real-world clinical validation. npj Digit. Med. 7, 46, 2024.

10. Huanbutta, K.; Burapapadh, K.; Kraisit, P.; et al. Artificial intelligence-driven pharmaceutical industry: A paradigm shift in drug discovery, formulation development, manufacturing, quality control, and post-market surveillance. Eur J Pharm Sci, 1;203:106938, 2024.

11. Yang, W.; Wei, Y.; Wei, H.  et al.  Survey on Explainable AI: From Approaches, Limitations and Applications Aspects. Hum-Cent Intell Syst, 3, 161–188, 2023.

12. Pun, F.W.; Ozerov, I.V.; Zhavoronkov, A. AI-powered therapeutic target discovery. Trends Pharmacol Sci, 44(9):561-572, 2023.

13. Meijer, C.; Uh, H.W.; El Bouhaddani, S. Digital Twins in Healthcare: Methodological Challenges and Opportunities. J Pers Med, 23;13(10):1522, 2023.

14. Mann, M.; Kumar, C.; Zeng, W.F.; Strauss, M.T. Artificial intelligence for proteomics and biomarker discovery. Cell Syst, 18;12(8):759-770, 2021.