ESAI: Ethically Safe AI (2023 - 2024)

Abstract

Increasingly, tasks which were formerly carried out by humans are now delegated to "intelligent" machines. However, the replication of human cognitive capacities in the form of machines means that the concept of risk must be continuously recalibrated. This has to do with the fact that the assumption of responsibility is handled differently in the case of humans than in machines, which do not make "mistakes", strictly speaking.

The initial focus of the research proposal is to create a common vocabulary which enables the use of concepts from ethics in the field of cognitive cyber-physical systems.

The second step will be to choose potential use cases of AI which the IKS is working on, in which risk-ethical considersations (for example) play a role. As part of this, an up-to-date summary of problems should be drawn up which is relevant to the applications which the IKS is working on.

Project information