Friedrich von Bohlen
Artificial Intelligence in Clinical Drug Development
Clinical development is slow, costly, and uncertain—often a decade from first-in-human to approval, with only about one in ten candidates succeeding. As biology reveals itself as a dynamic, multi-omic system, AI becomes the connective tissue: structuring heterogeneous data, modeling causality, and guiding decisions from diagnosis to trial design.
Molecular biology will be for medicine what mathematics is for physics. And AI will transform clinical drug development from a highly inefficient into an efficient process.
Drug programs fail less from lack of effort than from limited causal understanding. Phenotypes emerge from shifting molecular setups across organs and time; interventions ripple through pathways with intended and unintended effects. AI changes the substrate: integrating multi-omic, clinical, imaging, cohort, and longitudinal data; linking phenotypes, targets, and drugs; and supporting inbound curation, analytical modeling, and outbound decision support. The destination is precision—diagnoses tied to mechanism, trials designed around predicted responders, and safety signals anticipated rather than discovered late. The turning point, the author argues, will be regulatory acceptance of simulated trials with small confirmatory cohorts—shortening timelines, lowering costs, and reducing failures while clarifying who should receive which therapy under which conditions.
When Trials Learn Faster:
A New Logic for Development
What you’ll learn
-
How phase transitions, timelines, and costs shape risk across Phase I–III—and why overall success hovers near 10 percent
-
Why causal understanding across genomes, proteomes, metabolomes, and other “-omes” is central to precision therapy
-
How AI unifies heterogeneous sources—clinical, molecular, imaging, literature—into models that link phenotypes, targets, and drugs
-
Which development steps AI can accelerate: cohort selection, responder prediction, safety signal anticipation, and protocol design
-
What regulatory shift could unlock efficiency: simulated trials with few-patient confirmatory studies and clearer labeling for who should be treated under which conditions
Friedrich von Bohlen is co-founder and CEO of Molecular Health GmbH. He holds a diploma in biochemistry from the University of Zurich and a PhD in neurobiology from the Swiss Federal Institute of Technology (ETH). He is chairman of the board of Orexo AB and of InnoSource Ventures AG.
About the Author
Dr. Friedrich von Bohlen
Connect
