BAMKI

In view of the growing shortage of skilled workers and the potential impact on the performance of logistics chains and transport infrastructure, solutions are needed to support and relieve the burden on people in the context of automated mobility and logistics.
AI-based service robots are being deployed and demonstrated in various use cases within the logistics sector to relieve the burden on human resources. In particular, high safety and security requirements are placed on the AI components in the interactive deployment of service robots to ensure that people are not endangered when interacting with the robot.
Building on the Safety-by-Design methodology for AI-based components, the BAMKI project focuses on extending this methodology to incorporate Security-by-Design aspects. This enables AI-based components to be systematically protected against potential attacks from the outset. Safe and secure AI components are being developed to demonstrate the methodology's effectiveness.

Transferable solutions for safe human-AI automation

Another focus of the project is the safe design of human-AI cooperation. To this end, the HAICoSAFE (Human-AI-Robot Collaboration for Safe Automation – Framework and Evaluation Methods) framework is being developed. This is demonstrated through practical application examples in the transport sector, such as a cooperative multi-agent environment for central logistics processes (U-Shift) and AI systems for ship maintenance, in conjunction with the Security-by-Design methodology.
The results of the BAMKI project offer great potential for future projects, as the AI tools and simulations, as well as the HAICoSAFE framework, can be flexibly adapted to different contexts.

Contribution of the Institute for AI Safety and Security

As the lead institute, we are involved in various aspects of the project. Our responsibilities include the further development of the Safety-by-Design methodology and the development of secure AI components. We are also responsible for developing the HAICoSAFE framework. Use cases are planned and implemented at the institute’s own large-scale research facility, 'RAPID'. By developing and implementing transferable methods, AI tools and frameworks for safe human–AI cooperation, the institute is making a significant contribution to the design of trustworthy AI systems for use in the mobility and logistics sectors.