Can AI Improve Success Rates in Drug Development?

By Rasit Dinc

The pharmaceutical industry has long been plagued by the staggering costs and high failure rates associated with drug development. The journey from identifying a potential drug candidate to gaining regulatory approval is a long and arduous one, often taking over a decade and costing billions of dollars [2]. A significant portion of this cost is attributed to the high attrition rate of drugs in clinical trials, with estimates suggesting that approximately 90% of drugs that enter clinical trials fail to make it to market [3]. In recent years, artificial intelligence (AI) has emerged as a transformative technology with the potential to revolutionize this process and significantly improve the success rates of drug development.

The Role of AI in Enhancing Preclinical and Clinical Research

Artificial intelligence is already making significant inroads in the early stages of drug discovery and development. AI algorithms can analyze vast datasets of biological and chemical information to identify novel drug targets and design molecules with a higher probability of success. This is reflected in the remarkable success rates of AI-discovered drugs in early-stage clinical trials. A recent study found that AI-discovered molecules have an 80-90% success rate in Phase I trials, a substantial improvement over historical industry averages [1]. This high success rate is attributed to AI's ability to design molecules with optimized drug-like properties and safety profiles, thereby reducing the likelihood of failure due to toxicity or other adverse effects [1].

Beyond the preclinical stage, AI is also transforming the design and execution of clinical trials. AI-powered platforms can optimize patient recruitment by identifying eligible participants from electronic health records and other data sources. This not only accelerates the recruitment process but also helps to ensure that the trial population is representative of the intended patient population [2]. Furthermore, AI can be used to generate synthetic control arms, which are virtual patient groups created from real-world data. This can reduce the need for placebo groups in some trials, making them more ethical and efficient [2].

Challenges and the Road Ahead

Despite the promising results, the widespread adoption of AI in drug development is not without its challenges. One of the primary hurdles is the availability of high-quality data. AI models are only as good as the data they are trained on, and the pharmaceutical industry is often hampered by siloed and inconsistent datasets [3]. To unlock the full potential of AI, there needs to be a concerted effort to standardize and share data across the industry.

Another challenge is the so-called "black box" nature of some AI models, which can make it difficult to understand how they arrive at their conclusions. This lack of transparency can be a major obstacle in a highly regulated industry like pharmaceuticals, where it is crucial to be able to explain and justify every decision [3].

Despite these challenges, the future of AI in drug development looks bright. The US Food and Drug Administration (FDA) has already released draft guidelines for the use of AI in regulatory decision-making and has even developed its own large language model (LLM) to accelerate the review of clinical trial protocols [2]. As AI technology continues to mature and the industry becomes more adept at leveraging its power, we can expect to see even more significant improvements in the efficiency and success rates of drug development. The integration of AI into the drug development pipeline has the potential to bring new, life-saving therapies to patients faster and at a lower cost.

References

[1] Jayatunga, M. K., Ayers, M., Bruens, L., Jayanth, D., Meier, C., & van der Schaar, M. (2024). How successful are AI-discovered drugs in clinical trials? A first analysis and emerging lessons. Drug Discovery Today, 29(6), 104009. https://doi.org/10.1016/j.drudis.2024.104009

[2] Clinical trials gain intelligence. (2025). Nature Biotechnology, 43(7), 1017–1018. https://doi.org/10.1038/s41587-025-02754-1

[3] Sun, D., & Macedonia, C. (2025, January 7). Will AI revolutionize drug development? Researchers explain why it depends on how it’s used. Journal of Health Economics and Outcomes Research. Retrieved from https://jheor.org/post/2904-will-ai-revolutionize-drug-development-researchers-explain-why-it-depends-on-how-it-s-used