Collaborative Model-Based Process Assessment for trustworthy AI in Robotic Platforms

The use of robots in combination with artificial intelligence (AI) is a trend with the promises to relieve humans from difficult-, time consuming- or dangerous work. Intelligent robots aim to solve tasks more efficiently, safer or partly more stable. Independent of the domain-specific challenge, the configuration of both (a) the robot and (b) the AI currently requires expert knowledge in robot implementation, security and safety regulations, legal and ethical assessments and expertise in AI. In order to enable a co-creation of domain-specific solutions for robots with AI, we performed a laboratory survey – consisting of stakeholder interaction, literature research, proof-of-concept experiments using OMiLAB and prototypes using a Robot Laboratory – to elicit requirements for an assistant system that (i) simplifies and abstracts robot interaction, (ii) enables the co-creative assessment and approval of the robot configuration using AI, and (iii) ensures a reliable execution. A model-based approach has been elaborated in the national funded project complAI that demonstrates the key components of such an assistance system. The main concepts paving the way for a shift from research and innovation into real-world applications are discussed as an outlook.