Artificial intelligence is becoming central to discussions about the future of drug development, but a critical gap threatens to undermine its potential: most clinical trials simply don't collect enough detailed biological information about patients. While AI excels at finding patterns in large datasets, it cannot uncover information that was never gathered in the first place. Without richer patient data, even the most sophisticated AI algorithms will struggle to deliver on their promise to accelerate drug discovery and improve treatment outcomes.

What's Holding Back AI in Clinical Trials?

For decades, clinical trials have been designed with a narrow focus: determine whether a treatment is safe and effective. That objective remains essential, but this traditional approach leaves enormous blind spots. Trials typically collect only the information needed to meet regulatory endpoints, which means they often miss opportunities to understand why some patients respond to a therapy while others don't.

The problem runs deeper than incomplete data collection. Patients respond differently to treatment for multiple reasons, including genetics, age, ethnicity, environment, disease stage, and broader biological variation. Yet clinical trials rarely capture this diversity or investigate the underlying mechanisms that drive different responses. This gap is particularly problematic for precision medicine, which aims to match therapies more effectively to specific patient groups.

"AI probably pulls the efficiency and speed lever when it comes to data interpretation. Collecting data from the internal and external world, medical writing, predicting patient response and modelling trial designs are all areas of work that take a long time. But once the data is there, AI can really accelerate things in the future," said Graham Price, Head of Global Clinical Development at UCB.

Graham Price, Head of Global Clinical Development at UCB

How to Build Better Clinical Trials for AI Success

• Collect Richer Biological Data: New technologies now allow researchers to examine tissues, cells, and disease processes at unprecedented levels of detail, capturing information about gene expression, protein activity, molecular pathways, cellular interactions, and tissue architecture.
• Embrace Patient Diversity: Clinical trials must actively recruit and study diverse patient populations to understand how biological characteristics vary across different groups and how those variations affect treatment response.
• Treat Trials as Learning Opportunities: Rather than viewing clinical trials solely as mechanisms for testing drugs, they should be redesigned to generate knowledge about human biology, disease mechanisms, and why treatments work differently across populations.

Initiatives like the Smart Trials Hub at King's College London are already demonstrating this approach in practice. The Hub combines tissue samples, clinical information, advanced molecular profiling from diverse patient populations, and patient-derived models to create a deeper understanding of disease biology. When researchers generate high-dimensional datasets across diverse populations, they can identify the molecular features associated with treatment responses and understand why those responses vary between individuals.

Why Does This Matter for AI Development?

AI is exceptionally good at identifying patterns within large and complex datasets, but it cannot uncover information that has never been collected. If the underlying data fails to capture the biological diversity of patients or the mechanisms driving treatment response, even the most sophisticated AI algorithms will have limited value.

The good news is that our ability to generate this critical data has advanced dramatically in recent years. Patient-derived models and organoids can help researchers investigate how biological features influence treatment response. When AI and machine learning tools are applied to these richer datasets, they can integrate and interpret complex information, revealing patterns that would otherwise remain hidden.

According to experts, AI can help accelerate data interpretation and trial design in several ways. For example, AI models can analyze molecular structure and mechanisms of action, then help select patient populations where a therapy might work most effectively. This approach could increase the number of indications and patient types that can be treated with a single new medical entity.

What Does Success Look Like?

The next generation of clinical trials must maintain a focus on outcomes while also gathering and generating as much knowledge as possible. Success should not be measured solely by whether a study reaches its endpoint, but also by how much researchers learn about disease biology and patient response along the way. Increasing the amount of data retrieved from every patient will massively accelerate machine learning tools' ability to spot patterns and derive new hypotheses for validation.

Achieving this goal will require close partnership between industry and academia to effectively harness emerging innovations and secure data sharing to maximize the accuracy of AI models. The shift in thinking is particularly important in precision medicine, where the practical goal is understanding broader biological characteristics that define different patient groups and using that knowledge to match therapies more effectively.

AI will undoubtedly become an increasingly important part of clinical development, but better algorithms alone will not solve the challenges facing clinical research. The real opportunity lies in understanding where AI can provide true value in accelerating and improving clinical trial outcomes through biology-informed decisions, grounded in comprehensive patient data that reflects the full diversity of human biology.