Laboratories capable of running experiments, analysing the results and deciding what to test next are the focus of a new £150,000 initiative at King’s College London.
King’s Autonomous Labs (KAL) will bring together artificial intelligence, robotics, sensors and laboratory automation to explore the development of so-called ‘self-driving’ laboratories.
Rather than simply automating repetitive tasks, these systems could use data gathered during an experiment to determine the next step, allowing experiments to adapt as results emerge. Researchers would remain in control, defining the scope of projects and the safeguards under which autonomous systems operate.
The initiative will begin with a £150,000 internal funding programme supporting six-month proof-of-concept projects involving living systems. Their complexity and constantly changing nature make them a particularly demanding test for autonomous technology, which must be capable of responding to unexpected results.
Advertisement
The projects are expected to produce experimental data, prototypes and new processes, with the resulting technologies and infrastructure eventually applied across a wider range of scientific disciplines.
‘By connecting these strengths, we aim to accelerate scientific discovery, support interdisciplinary collaboration and establish King’s as a leader in the development of autonomous laboratories,’ said Professor Deepak Srivastava, Professor of Molecular Neuroscience at King’s.
Autonomous laboratories are part of a growing effort to reduce the time between forming a scientific question and testing it. By connecting instruments and sensors directly with AI systems, researchers could potentially run and refine experiments with far less manual intervention.
King’s said the technology is intended to extend what scientists and laboratory staff can test, rather than replace their expertise. KAL will also develop governance around the safe and responsible use of autonomous experimental systems.
The initiative forms part of the King’s Institute for Artificial Intelligence and its ‘Doing Science Well with AI’ programme.
Advertisement