Communicating AI: Knowledge Structures and Sensemaking Tools for Critical Engagement

As AI systems become ubiquitous in many areas of life, including education and work, it is essential that they are better understood by a variety of stakeholders. This includes different types of end-users in different arenas, such as AI engineers that create those systems, managers that deal with the embedding of AI systems in their organisations, educators or trainers in various domains, or non-expert audiences. For these, technical knowledge about AI technologies and related regulations (e.g., EU AI Act) is increasingly important. Such knowledge also includes critical reflection on and healthy scepticism towards the outputs of AI systems, including knowledge of limitations and biases.

A key challenge lies in how to efficiently and effectively equip these stakeholders to acquire (subsets of) the knowledge that they require for performing their various tasks. In this thesis, the aim is to explore various methods for AI knowledge communication and co-production relying on formal knowledge representation (to represent different expertise levels) and advanced visual interactions. We will test the effectiveness of such methods with an explicit focus on critical evaluation. Through this we aim to foster critical interaction with AI systems and a better understanding of human agency when working with AI.

Marta Sabou – Department of Information Systems and Operations Management, WU Wien

Laura Koesten – Faculty of Computer Science, University of Vienna

Hajo Boomgaarden – Department of Communication, University of Vienna