Making artificial intelligence understandable: Constructing explanation processes

Sifting through job applications, analyzing X-ray images, suggesting a new track list--interaction between humans and machines has become an integral part of modern life. The basis for these artificial intelligence (AI) processes is algorithmic decision-making. However, as these are generally difficult to understand, they often prove less useful than anticipated. Researchers at Paderborn and Bielefeld University are hoping to change this, and are discussing how the explainability of artificial intelligence can be improved and adapted to the needs of human users. Their work has recently been published in the respected journal IEEE Transactions on Cognitive and Developmental Systems.