Motivation

The research programme focuses on the use of AI assistants in production systems and supply chains. AI assistants can support human decision-makers by supplementing or even replacing and automating domain knowledge and expert decisions. Improved planning and control in dynamic scenarios, where manual control cannot do justice to the complex interrelationships, can lead to increased efficiency and cost reductions. In addition, increased processing capacity can increase transparency in production and supply chains and lead to competitive advantages.

The challenge for future production systems and supply chains in the direction of Industry 4.0 is that they must be dynamically reconfigured at runtime, for example to be able to produce and deliver highly customised products (up to batch size 1). The project focuses in particular on how to ensure the reliability of AI assistants in critical application areas. Critical areas include all applications that have a high risk of harming people or causing major material damage. Examples of this would be the use of AI for the dynamic reconfiguration of the production plant / supply chain or the use of AI-supported cobots or driverless transport systems. As part of the five doctoral programmes, processes, methods and tools for the development and use of reliable AI assistants are to be created.

Activities

Of the five doctoral positions, two are based at RPTU and three at Mainz University of Applied Sciences. Two of the latter are being carried out at the Department of Economics, one at i3mainz under the supervision of Jens Heidrich.

This thesis will focus on the testing of AI in critical applications, with the aim of creating a test framework for language models.