The use of Artificial Intelligence has entered many application areas, mainly thanks to the wide availability of Deep Learning approaches together with their models and tools. Many of these concepts are tangible: for example, when we think of the new results of image analysis, where photos can be automatically labelled with content elements, objects on images can be uniquely recognised or the digital voice assistant recognises our language and keeps shopping lists for us.
Even with these applications, it is important for both technical experts and users to understand how the neural network came to its decision and what factors influenced this decision. In many other areas, AI technologies will be integrated into complex socio-technical systems in the future and clearly influence our lives, for example, supporting automated or semi-automated decisions and changing work processes.
In a future in which we humans will be empowered with and through digital technologies to successfully solve complex tasks and problems and thus benefit from technology in our everyday and professional lives, it is elementary to create transparency about the capabilities and limitations of AI-supported systems and to offer explanations of AI decision-making - especially also in order to be able to provide targeted support for people with very different qualification levels and abilities.
Aljuneidi, Saja and Heuten, Wilko and Tepe, Markus and Boll, Susanne; Proceedings of the 2023 ACM Conference on Information Technology for Social Good; 2023
Schultze, Sven and Withöft, Ani and Abdenebaoui, Larbi and Boll, Susanne; Proceedings of the 2023 ACM International Conference on Multimedia Retrieval; Juni / 2023
Withöft, Ani and Abdenebaoui, Larbi and Boll, Susanne; International Conference on Multimedia Modeling; 2022