1.7 million books to have AI explain why certain molecules come together

The European Research Council is funding a facility in Durham that combines a robotic laboratory and explainable AI, with the aim of reducing the discovery of functional molecules from several decades to a few weeks.

The chemist Filip Szczypiński, from Durham University, received a start-up grant from the European Research Council worth 1.7 million pounds, or about 2.1 million euros. His project aims to establish the design rules that govern molecular self-assembly—the way molecules recognize and spontaneously organize themselves.

The device is based on two complementary bricks. The robotic laboratory automatically chained thousands of assembly reactions, producing the volume of experimental data that would never be reached by manual work. The explainable AI then intervenes to clear the regularities and, above all, to return the reasons why some assemblages succeed: an opaque prediction would not help a chemist to conceive the next molecule.

The applications are sensors capable of detecting dangerous diseases and substances, as well as catalysts that make industrial and pharmaceutical processes faster and less polluting. The team argues that such a loop could reduce the development of functional molecules from several decades to a few weeks, where the test error still dominates.