The Swedish Chemicals Agency (KEMI) has developed an AI model that uses data from patent applications to recognize chemical risks as early as the product development phase. The developers of the model want to ensure toxin-free circulation according to the requirements of the Green Deal.
During development of the model, the agency collaborated with the Swedish Patent and Registration Office and RISE, the state research institute.
How well the method functions was evaluated as part of a case study in which the program was to identify polychlorinated biphenols (PCBs) in patents. The method worked in 96% of the cases. Patents provide information on the use of chemicals in products before the items go on the market, so that this information enables proactive regulation of hazardous chemicals.
The next step will examine how to improve the quality of the data through a combination of various AI methods and how well such methods would be appropriate for other application areas.
Do you need any support with the registration of substances? Please contact us under email@example.com.