AI can compress drug discovery timelines, unlock real-world health gains

Amazon Web Services (AWS), Boston Consulting Group (BCG), and Merck have launched an artificial-intelligence (AI) platform to improve clinical-trial site selection. While AI is rapidly compressing early-stage drug-discovery timelines, from months or even years to weeks, clinical trials remain the slowest, costliest, and most failure-prone stage of the pipeline. AWS’ new AI-powered “Amazon Bio Discovery” platform is built to accelerate early-stage drug discovery by enabling scientists to execute complex computational workflows without needing to code. Academic studies suggest that AI-enabled methods can help overcome core inefficiencies in clinical-trial operations, potentially speeding up drug-development milestones, such as reducing database lock timelines by around 33 per cent, while ensuring regulatory compliance and lowering operational costs. At the same time, machine learning has improved programmatic edit checks by leveraging historical data to finetune thresholds and minimise false positives.

 

However, translating this potential into systemic gain is far from straightforward. The first challenge is data fragmentation. Health data remains in silos across hospitals, insurers, and geographies, often locked behind incompatible systems. Even large pharmaceutical firms struggle with integration. Without interoperable data infrastructure, AI models risk being trained on incomplete or biased datasets. Second, there is regulatory uncertainty. While regulators such as the United States Food and Drug Administration and European Medicines Agency are open to AI-assisted methods, a lack of alignment complicates regulatory interpretation. Third, algorithmic bias poses a serious risk. If AI models are trained predominantly on specific demographic groups, they may reinforce existing inequities in trial participation, especially for populations in the Global South. Finally, there is the issue of unequal global access. Advanced AI platforms are being developed largely by global technology firms and deployed in well-resourced research ecosystems. Without deliberate policy intervention, low- and middle-income countries, despite bearing a large disease burden, may remain peripheral to the clinical trial landscape.

 

Addressing these challenges requires coordinated action. Governments must prioritise interoperable digital health infrastructure, enabling secure data sharing across institutions while protecting privacy. Investment in national health-data grids, standardised formats, and federated learning models can help unlock the full value of real-world data. Equally important is the development of robust AI governance frameworks. This includes standards for model validation, bias audits, and explainability, as well as clear lines of accountability when AI systems influence clinical decisions. Public trust will be critical to adoption. The pharmaceutical industry, for its part, must embrace inclusive trial design. AI tools should be explicitly calibrated to improve diversity in recruitment, expanding trial networks beyond traditional urban and high-income settings. Partnership with local health systems can help bridge this gap. Initiatives like the AWS-BCG-Merck platform signal the path ahead: Aligning data systems, regulation, and equity is the way forward for AI to truly unlock faster, more inclusive clinical trials and ensure that innovation translates into accessible, real-world health gain.