The letter below was published in The Economist in January 2026 in response to its Science & Technology feature “An AI revolution in drugmaking is under way”, which examined how artificial intelligence is reshaping drug discovery and development.
"You highlight how artificial intelligence is shortening the path to new drugs (“The imagination factory”, January 10th). However, the revolution has yet to reach cancer patients. Oncology trials still operate as a lottery where patients exhaust their health, and in some cases life savings, on therapies that were never likely to work for their specific biology.
The synthetic patient, or digital twin, technology you describe represents progress, yet it suffers from a fundamental limitation: failing to account for an individual’s biological complexity. Current approaches match patients by demographics and medical history, essentially, high-tech filters to identify historical analogues. But two patients with identical clinical profiles often respond to the same drug in radically different ways. This is because clinical characteristics tell us about the patient, not their tumour; their history, not their future.
By modelling the molecular machinery of individual tumours rather than historical controls, AI can more effectively predict which therapies will work before the first dose is administered, whether in trials or in standard care. This approach requires a collective commitment from clinicians, researchers, funders and regulators to prioritise mechanistic understanding over demographic matching."
Dr Irina Babina
Chief Executive
Concr
Asghar, U.S., Chung, C. Application of digital twins for personalized oncology. Nat Rev Cancer 25, 823–825 (2025). https://doi.org/10.1038/s41568-025-00850-7
Back
LinkTree