AI for the German Public Sector: A Technical Concept for Digital Sovereignty, EU AI Act Compliance and Enhanced Productivity

In this work, a concept is presented for enabling German governmental authorities to make use of artifi cial intelligence (AI) to increase each civil servant’s productivity. This boost in productivity is achieved by assisting administrative processes with AI and is necessitated by the demographic change.

This concept includes every layer of implementation, from specific hardware tailored to the needs of the fi nal software solution to the political guardrails of digital sovereignty and the EU AI Act.

The core idea of the fi nal solution is to ensure high adaptability in both software and hardware to accommodate various use cases and changing future circumstances. The solution combines a modular software architecture with an agent-based framework built on maintained, tested, and proven open source projects, which are considered best practice in sovereign software solutions. To support this, the hardware architecture relies on technology with open standards, minimizing critical manufacturer dependencies to strengthen digital sovereignty at this level as well. These approaches guarantee the highest level of digital sovereignty currently achievable in Europe and ensure EU AI Act compliance by design. 

The software solution can be understood as the interplay of different specialized modules. The hardware leverages the cloud-native, open source foundation of Kubernetes to provide a scalable and future-proof platform for operating AI efficiently. Furthermore, the modular architecture en-
ables the continuous evolution of the platform, making it fully adaptable to every new circumstance.

In summary, this work outlines a concept to build a digital infrastructure from scratch to enable the use of AI as a tool within German governmental authorities, while maintaining digital sovereignty and complying with the EU AI Act. The core motivation of this work is to empower decision-makers at various levels of public administration to develop the best possible solutions for the public good.