Swedes lead in toxic chemical detection by using revolutionary AI method

A pioneering approach for evaluating harmful substances, relying solely on molecular structure, has been invented by scientists at Chalmers University of Technology and the University of Gothenburg in Sweden. This advancement can potentially improve chemical control and understanding in society, reducing animal testing requirements.

The use of many chemicals is commonplace in various facets of modern life, and they can be found in everything from household items to industrial operations. Upon entering the water supply and natural habitats, such chemicals can threaten humans and various species.

For instance, Per- and Polyfluoroalkyl Substances (PFAS), a problematic group of substances, were discovered in elevated concentrations in groundwater and drinking water sources. PFAS is found in firefighting foam and other consumer products.

Despite stringent chemical regulations that frequently necessitate time-consuming animal testing to establish safety, negative impacts on humans and the environment persist. The EU uses over two million animals each year to fulfill various regulatory requirements. Moreover, with new chemicals continually emerging, determining which ones require restriction presents a significant challenge. Thus, it’s crucial to identify toxic substances early and reduce reliance on animal tests.

At Chalmers University of Technology and the University of Gothenburg, Professor Erik Kristiansson discusses how assessing the toxicity of approximately 100,000 commercial chemicals through conventional techniques like animal testing is not feasible. He proposes an alternate method as a viable solution. Kristiansson and his fellow researchers have made their work publicly available for others to build upon.

Enhanced accuracy and applicability

Previous computational tools for toxic chemical detection existed, but they had limited applicability domains or low accuracy, preventing significant replacement of lab tests.

The type of AI we use is based on advanced deep learning methods

Erik Kristiansson

The novel AI method developed by Swedish researchers demonstrates superiority over previous computational tools used for toxic chemical detection. It not only broadens its scope of application but also guarantees greater levels of accuracy, making it a promising replacement for laboratory tests in assessing the toxicity of chemicals.