Researchers at the University of British Columbia have trained computers to predict the next designer drugs before they even appear.

New types of designer drugs are created every year. Design (from the English. To design – to design, develop) refers to substances that are developed in order to circumvent prohibitive legislative norms. Due to the lengthy process of identification and classification, they pose a particular danger, since they are legal up to a certain point. Revealing so-called “legal highs” in seized pills or powders can take months, during which thousands of people may have already used the new designer drug.

According to the University of British Columbia, the new development is already helping law enforcement agencies around the world reduce the time it takes to identify designer drugs from months to days, which is critical in the race to identify and regulate new versions of dangerous psychoactive drugs.

Scientists used a database of known psychoactive substances provided by forensic laboratories around the world to train artificial intelligence. They used a neural network that studied the chemical structures of drugs. Through training, the model created about 8.9 million potential designer drugs.

The resulting molecules were compared with 196 psychoactive substances that appeared on the illegal market after training the model. It turned out that she predicted 90% of new designer drugs. The research results are published by the journal Nature Machine Intelligence.