King's College London

A team from the Department of Engineering, led by Dr Miao Guo, has been awarded a £500,000 grant from ARIA (Advanced Research + Invention Agency) to develop an AI-enabled autonomous lab that could transform by-products from the agro-food industry [KN1] into high quality protein, helping address global challenges in hunger and sustainability.

Working with co-Investigators Professors Yansha Deng and Chris Lorenz, the project brings together their expertise in molecular communication, edge AI and molecular dynamics, with Miao’s work in AI-driven biochemical process optimisation. Together, the team aims to build an AI system capable of performing the full scientific process autonomously, from hypothesis generation, automated experimentation design and execution, to optimising how these proteins are developed.

Autonomous laboratories for scientific experiment are increasingly being explored worldwide to accelerate scientific discovery, with the UK government now calling for teams to build the country’s capabilities in the field.

Building on this momentum, the ‘Towards a Self-Reflective AI Scientist for Autonomous Sustainable Microbial Protein Biomanufacturing’ project envisions a new generation of self-driving labs that merge digital AI with chemical and molecular intelligence.

By combining AI scientist with electronic-free molecular signal processing, this project will create energy-efficient autonomous lab that can generate, test and refine scientific hypothesis – pushing the sustainable production of important materials through biological means beyond the capabilities of conventional human-led experimentation.

Reflecting on the potential of autonomous science, Dr Guo, Senior Lecturer in Engineering, said “Experimental science is essential in pushing the boundaries of human knowledge, but transitional labs are often costly, inefficient and constrained by manually co-ordinating a project all the way from discovery to implementation. Autonomous labs offer a fundamentally new way to rethink how science is done”.

This programme brings together AI and automated hardware to create a self-reflective AI scientist, a system that does not just assist researchers, but actively reasons like one.

Operating around the clock, the AI scientist coordinates and directs autonomous robots and analytical devices to run hundreds of experiments and real-time measurements on how microbes synthesise complex carbon sources into protein for food. A virtual laboratory will then be layered onto these experiments, exploring thousands of different conditions and parameters to improve the process.

By learning from large volumes of data in real time and rapidly refining experimental strategies, the system removes human-centred bottlenecks, accelerates discovery, reduces costs, and enables scientific research to scale far beyond the limit of a traditional laboratory.

Looking ahead, the team believes the same AI Scientist framework could extend beyond biomanufacturing. By creating laboratory environments which more closely mimic human biology, the AI scientist framework could one day help accelerate medical research, including the development of virtual and physical models of complex systems such as the human gut microbiome.