Robots found to perform chemical research ‘faster’ than humans

AI-DRIVEN robots can carry out complex chemistry experiments “faster” than humans, according to a University of Liverpool study.

A team of researchers from the UK university have designed mobile robots which can perform exploratory chemistry tasks, including experiments, product analysis, and decision-making.

The researchers believe the intelligent robots’ speed could make research a less time-consuming and expensive process.

Sriram Vijayakrishnan, a current postdoctoral researcher with the Department of Chemistry who led the study, said: “When I did my PhD, I did many of the chemical reactions by hand. Often, collecting and figuring out the analytical data took just as long as setting up the experiments.

She added: “We tackled this here by building an AI logic for the robots. This processes analytical datasets to make an autonomous decision.”

Tough decisions

Robotics and automation have been used as lab research aids for decades, with robots designed to use small-scale objects, including test tubes and pharmaceutical tablets.

Similar AI robot studies have been trialled this year, with the University of Amsterdam using AI robots in flow chemistry. The robots from the study were able to collect positive and negative data from experiments.

The Liverpool researchers took this a step further, designing the robots to make decisions based on data.

Using two AI-driven robots, the researchers addressed three areas of chemical synthesis, namely structural diversification chemistry, supramolecular host-guest chemistry, and photochemical synthesis.

The study found that using AI logic, the robots could make the same decisions as a synthetic chemist could in these areas, but where it would take chemists hours, the robots made decisions in “the blink of an eye”.

Industry application

The team plan to use the robotic technology to develop chemical research relevant to pharmaceutical drug synthesis, and materials for carbon capture.

Andrew Cooper, a professor from the university, said: “There is also huge scope to expand the contextual understanding of the AI, for example by using large language models to link it directly to relevant scientific literature.”