AI Supports Intervention Detection (and Classification) in Aseptic Core

Fresenius Kabi developed an artificial intelligence (AI) supported detection system which is capable to detect and classify interventions performed in the aseptic core of an isolator. Due to the current absence of detailled regulatory guidance targeting validation of AI powered systems in the GMP environment, Fresenius Kabi also developed a validation approach. This new approach does not rely on black box models but help users to understand how AI decided, where its limitations are but also which disruptive potential it has on the way we operate aseptic installations.