From Failure to Innovation: How AI is Reshaping Drug Trials in Pharma

The pharmaceutical industry maintains a persistent 90% failure rate in investigational drug trials despite implementing various development strategies. ^[1-3] The high drug development expenses and extended time requirements have created an immediate need for new solutions. The pharmaceutical industry now uses Artificial Intelligence (AI), as a transformative technology, which addresses multiple problems that affect traditional drug discovery and development methods^{. [1, 4]}

The Scale Throughout the Industry

The pharmaceutical industry faces a 90% failure rate for drug candidates in clinical trials because their therapeutic effects are insufficient or their adverse effects prove too severe.^[1]  The clinical development stage carries substantial financial risks, as 90% of candidates fail to progress while R&D expenses account for up to 60% of total costs due to project termination.^{[2, 5]} The failure rate has shown no improvement over the years. The success rate for compounds entering Phase I trials remains approximately 10% according to current estimates, consistent with research findings from the past two decades.^{[4, 5]} These persistent failure rates indicate that conventional drug development methods have exhausted their potential, so new fundamental approaches need to be implemented. ^[1-3]

Understanding Why Drug Development Fails

1. Lack of Efficacy as the Primary Culprit

The main reason for failure of drugs in clinical development is their inability to effectively treat their designated disease indications. ^{[1, 3]}  These high failure rates indicate fundamental problems in the current methods of target selection and molecule designing and translating preclinical results into human therapeutic applications.^[4]

2. The Preclinical Translation Problem

Inadequate preclinical models combined with high false discovery rates in preclinical research leads to high failure rates in drug development.^[3] The industry faces its biggest challenge because laboratory results fail to translate effectively into human biological systems.^{[1, 4]}

3. Therapeutic Area Challenges

Neurological drug development faces the highest failure rates because of numerous late-stage and mid-stage clinical trial failures. The pharmaceutical industry has experienced numerous high-profile failures in difficult-to-treat conditions which include Alzheimer’s disease and schizophrenia and herpes simplex virus.^{[4, 6]}

The AI Revolution in Drug Discovery

Market Growth and Investment

The AI drug discovery market has grown rapidly with the AI market expanding from USD 200 million in 2015 to USD 700 million in 2018. Analysts predict it will reach USD 5 billion by 2024. This fast market growth demonstrates both the rising AI technology funding and the understanding of pharmaceutical companies of AI’s ability to revolutionize drug development. ^[7]

Economic Impact Projections

According to Morgan Stanley, the implementation of AI and machine learning in early-stage drug development could lead to 50 new therapeutic drugs within ten years which would create a $50 billion market opportunity.^[8]

Real-World Applications and Results

1. Exscientia: Pioneering AI-Designed Drugs

Exscientia is a leading organization demonstrating some of the most advanced  applications of AI in drug discovery.^{[8, 9]} Its AI-driven design process has reduced the typical development period from 4.5 years to 12-15 months.^{[10, 11]} Notably, Exscientia created a potent and specific immunomodulatory drug candidate after synthesizing its 150^th  molecule within 11 months.^[12] To further enhance its platform, the company has partnered with Amazon Web Services (AWS) to integrate AI and machine learning for end-to-end drug development automation using generative AI models.^[9]

2. Clinical Pipeline Status

According to Biopharma Trend, AI-driven drug discovery has already advanced to human trials, with 31 drugs in Phase II/III testing, 17 in Phase I, five in Phase I/II, and nine ongoing human patient trials involving AI-designed compounds.^[13]

3. Recent Developments and Challenges

Despite notable progress, challenges remain. For example, Insilico released Phase 2a trial results showing no statistically drug efficacy for their first end-to-end AI-generated drug.^{[10, 14]}  Such results show that while AI enhances drug discovery, it still cannot ensure clinical success because of the persisting fundamental biological and clinical obstacles^{.[10, 13, 14]}

Technology Behind AI Drug Discovery

1. Platform Integration

Exscientia’s platform combines generative AI models with robotic lab automation to accelerate the development of high-quality drug candidates through faster and less expensive processes.^[9] Across the industry, a key trend has emerged of merging AI systems with laboratory equipment to harness computational strength and experimental testing capabilities together.^[9,10]

2. Broader Industry Adoption

The pharmaceutical and biotechnology industries are increasingly embracing AI as the next major breakthrough in drug development, a perspective supported by Insilico and other leading companies driving this trend.^{[7, 9]}

Current Limitations and Realistic Expectations

1. Mixed Results

AI technology benefits the drug discovery process, but the technology remains at an early stage of development. Thus, it faces skepticism from investors and industry professionals alike, who recognize both its potential and boundaries, according to 2024 media reports.^[10]

2. The Translation Challenge Persists

AI technology accelerates drug development, yet scientists still face the challenge of translating laboratory results into human clinical success. The clinical trial results of AI-generated drugs show that better computational drug design methods do not directly lead to enhanced patient treatment outcomes.^[13]

The Path Forward

1. Integration Rather Than Revolution

AI technology delivers its greatest value through enhanced operational efficiency, shorter development times, and generating superior drug candidates for clinical testing —rather than complete replacement of traditional drug discovery methods.^[1] Exscientia’s AI-guided design method has proven successful in attracting investors and producing revenue which demonstrates its commercial potential.^[11]

2. Focus on Specific Applications

AI proves most effective when applied to well-defined drug development problems instead of trying to solve all challenges at once. The main areas where AI shows promise in drug development^[7,8,12] include:

• Accelerating molecular design and optimization processes
• Improving the identification and validation of drug targets
• Enhancing patient stratification methods to identify the most suitable candidates for clinical trials
• Reducing the time and cost of preclinical development

Conclusion

The pharmaceutical industry faces its most critical challenge with its approximately 90% failure rate in all clinical trials which is one of the biggest obstacles in contemporary medical practice. Despite implementing various successful strategies, this persistently high rate highlights the need for fundamental changes in drug development approaches.

AI technology shows promise by shortening drug development timelines from years-to-months while generating novel drug candidates more efficiently. The AI technology delivers its best value when used to solve particular defined problems in drug discovery instead of trying to replace conventional methods entirely.

The implementation of AI in early-stage drug development can lead to more compounds reaching human clinical trials. However, the recent inconsistent results of trials indicate that AI technology alone cannot guarantee successful outcomes. The field will achieve its most enduring progress through realistic expectations and sustained funding of AI capabilities which will decrease industry-wide high failure rates and create better treatments for patients.

References

1. American Society for Biochemistry and Molecular Biology. (2022, March 12). 90% of drugs fail clinical trials. https://www.asbmb.org/asbmb-today/opinions/031222/90-of-drugs-fail-clinical-trials
2. Graaf, P. H. van der. (2022). Probability of Success in Drug Development. Clinical Pharmacology & Therapeutics, Wiley Online Library. https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.2568
3. King, K. L., et al. (2019). Improving the odds of drug development success through human genomics: modelling study. Scientific Reports, Nature. https://www.nature.com/articles/s41598-019-54849-w
4. Sun, D., Gao, W., Hu, H., & Zhou, S. (2022). Why 90% of clinical drug development fails and how to improve it? https://pubmed.ncbi.nlm.nih.gov/35865092/
5. Wong, C. H., Siah, K. W., & Lo, A. W. (2019). Estimation of clinical trial success rates and related parameters. Biostatistics, Oxford Academic. https://academic.oup.com/biostatistics/article/20/2/273/4817524
6. (2024, December 16). 5 Clinical Assets That Flopped in 2024. https://www.biospace.com/drug-development/5-clinical-assets-that-flopped-in-2024
7. Integrating Artificial Intelligence for Drug Discovery in the Context of Revolutionizing Drug Delivery. PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC10890405/
8. Petrie-Flom Center, Harvard Law School. (2023, March 20). How Artificial Intelligence is Revolutionizing Drug Discovery. https://petrieflom.law.harvard.edu/2023/03/20/how-artificial-intelligence-is-revolutionizing-drug-discovery/
9. (2024). Exscientia Launches AWS AI-powered Platform to Advance Drug Discovery. https://investors.exscientia.ai/press-releases/press-release-details/2024/Exscientia-Launches-AWS-AI-powered-Platform-to-Advance-Drug-Discovery/default.aspx
10. MIT Technology Review. (2024, March 20). An AI-driven “factory of drugs” claims to have hit a big milestone. https://www.technologyreview.com/2024/03/20/1089939/a-wave-of-drugs-dreamed-up-by-ai-is-on-its-way/
11. Exscientia: a clinical pipeline for AI-designed drug candidates. https://www.ukri.org/who-we-are/how-we-are-doing/research-outcomes-and-impact/bbsrc/exscientia-a-clinical-pipeline-for-ai-designed-drug-candidates/
12. Artificial Intelligence (AI) Applications in Drug Discovery and Drug Delivery: Revolutionizing Personalized Medicine. PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC11510778/
13. Progress, Pitfalls, and Impact of AI‐Driven Clinical Trials. PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC11924158/
14. STAT News. (2024, December 3). Is this the beginning of the AI-in-drug-discovery era, or the beginning of the end? https://www.statnews.com/2024/12/03/ai-drug-discovery-investors-insilico-recursion/

Authors:

Raghuraj Puthige, PhD., eMDP
Function Head, Medical Communications – Enago Life Sciences
Connect with Raghuraj on LinkedIn

Sweaksha Langoo (MSc. Molecular Biology and Biochemistry)
Scientific Writer – Enago Life Sciences
Connect with Sweaksha on LinkedIn