James Ladyman and Emanuele Ratti have been awarded a 1.1 million pound research grant from AHRC to investigate the epistemic impact of AI on chemistry.
In collaboration with the University of Durham, it will combine conceptual methodologies from philosophy of science with an ethnographic investigation of chemical practice in the age of AI.
This project will offer the first comprehensive and scientifically grounded study of the relation between AI and chemistry, by investigating the impacts of the use of machine learning on prediction, explanation and understanding in chemistry. The specific aim of the project is to demystify the role of AI in chemical practice by examining in detail how AI is used in chemistry to predict new chemical substances, and to discover and explain novel reaction mechanisms. This will be achieved by pursuing three objectives. The first is to analyse the predictive power of AI for the identification of novel forms of matter. The second objective is to analyse the use of machine learning tools in the investigation of new reaction mechanisms (the pathways through which chemical reactions happen that involve catalysts and intermediate products, as well as the reagents and products of the reaction as a whole). The third objective is to evaluate how machine learning will affect scientific understanding in chemistry, given that they may model chemical phenomena using forms of representation that humans cannot comprehend.
The methodology will combine conceptual and empirical inquiry, integrating rigorous philosophical analysis with ethnographic methods (including semi-structured interviews and embedded observations), informed by research on how key developments in the history of chemistry, such as chemical atomism, the Periodic Table of the elements, and theories of structure and mechanism have transformed chemical prediction, explanation and understanding. The project will draw upon and impact general philosophy of science, and it will provide a model that can be adapted to evaluate the role of AI in fields close to chemistry for which the impact of AI has also not been properly studied, such as quantum physics and biochemistry.
This novel project will be transformative for the disciplines of philosophy of chemistry, philosophy of AI and philosophy of science, and will have impact beyond academia. Given the far-reaching impact of AI on our cultural perceptions of science and its role in societal change, this project will also benefit society by offering a general blueprint for understanding how AI’s potential and the hype about it are connected to scientific progress, and prediction, explanation and understanding. It will inform how chemists, and scientists from other disciplines view the role of AI in their practice, and counteract misconceptions and unrealistic views about the impact AI can have on science.