1st project phase (2020-2022)

Within the framework of the basic-funded in-house research, the project addresses research questions that contribute in the context of program-oriented funding to the profile development of ITAS in the Helmholtz Research Field Information in the thematic area “Learning Systems”.

The research focuses primarily on the development and framing of a context-specific understanding of processes, methods, and consequences of learning systems as “enabling technologies” and also addresses possibilities for a responsible handling of potential risks in various fields of application. Since the overarching problem of trust in information systems is also affected in automated applications and decision-making systems, research will focus in particular on relevant questions concerning societal trust in learning systems. This will build on the project members’ previous and parallel work on risk research, autonomous driving, automation of the world of work, adversarial AI, explainable AI, and the philosophy of science in computational sciences, which will be further developed or made fruitful for application within the thematic field. One of the project’s important tasks is to bring together the previously unconnected work on this topic at ITAS and to develop a common perspective. This also includes organizing the transfer of knowledge across project and research groups and intensifying the cooperation with KIT computer science institutes.

2nd project phase (2023-2025)

During the first project phase, the assumption was confirmed that the project focus on societal trust in learning systems is an overarching topic that is increasingly gaining societal (and political) relevance. Trust is understood as an important resource not only in the AI discourse, but in the digitalization discourse in general.

In the second project phase, the conceptual work will be continued and further specified. The focus will be on:

1. Trust in regulatory institutions and its interrelation with trust in AI: trust in AI must be understood as a socio-technical phenomenon that is influenced by multiple actors and political dynamics across technologies. These include, in particular, current efforts to regulate AI at the EU level. However, this process still lacks conceptual acuity and empirical knowledge.
2. The relevance of trust in dealing with mis- and disinformation and its impact on the formation of societal and political opinion: artificial intelligence already plays an important role in this regard and will become even more important.
3. The role of trust in AI-driven automation processes: With regard to the high context specificity of artificial intelligence, constellations of trust in AI will be examined in selected contexts in which a great potential for social change can be expected, e.g., the world of work, mobility, energy.