As Germany aims to implement AI in public administration, Rhineland-Palatinate's Interior Ministry is taking concrete steps with a new AI unit and extensive training programs. State Secretary of the Interior Daniel Stich explains the state's strategy in an interview.
The German federal government aims to leverage artificial intelligence (AI) for the public sector through its “National Data Strategy.” However, Germany's federal states are the key players in implementing these plans. This particularly affects local and state administrations as well as law enforcement agencies - areas where citizens directly interact with government services and where the impact of AI use will be most noticeable.
As part of the "AI-Humanitas: Enabling AI-Skills" initiative, Dr. Wolfgang König interviewed Daniel Stich, State Secretary of the Interior for Rhineland-Palatinate.
The Decoder: Mr. Stich, as State Secretary in the Ministry of the Interior in Rhineland-Palatinate, what are the three most important lessons you've learned about AI in government and administration?
State Secretary Stich: First, it's important to demystify AI. It's not some alien force, but a tool that can support government and administration in many specific use cases. Second, we urgently need to build expertise within our administration. Staff members need to be excited about new technologies and recognize the opportunities for their specific tasks. Third, acceptance is crucial. People - whether citizens or employees - need to understand how AI works and how it helps. That's the only way to build trust.
The Decoder: The federal government wants to enable AI in administration. Is there a common position among Germany's interior ministers on this?
State Secretary Stich: AI was a major topic at the last Conference of Interior Ministers in December 2024 - particularly regarding the planned EU regulation and the use of AI in law enforcement. The discussion focused on how we can work together across state borders while complying with legal and ethical requirements from the EU regulation. The goal is to equip our law enforcement agencies with modern capabilities.
The Decoder: Two main areas stand out regarding AI: infrastructure and employee training. Where is AI technology already being used successfully, and what additional measures are planned?
State Secretary Stich: In the Ministry of the Interior, we've taken steps by establishing an AI staff unit at the beginning of the year, reporting directly to senior management. It's meant to be not just a think tank for the entire ministry, but also a “do tank” that identifies practical applications for our real-world challenges.
Additionally, our staff unit will serve an important networking function, bringing together existing approaches and best practices from various agencies and promoting exchange. We already have good initiatives, for example in law enforcement. The Rhineland-Palatinate police currently test or use AI to combat CSAM, process language and translations, and improve traffic safety work. We continuously identify new potential uses and evaluate their technical feasibility and legal compliance. We also research and develop targeted solutions in collaboration with academia and industry.
The Decoder: How are government and administration employees being trained to use AI in their daily work?
State Secretary Stich: For AI to truly succeed in administration, employees need proper preparation. There are already several nationwide initiatives, such as training programs and pilot projects where employees work with AI tools and provide feedback. In Rhineland-Palatinate, we've launched a training initiative that specifically incorporates AI and digitalization into basic and advanced training. It's important that we regularly evaluate whether the training is effective and adapt it to new technologies. Ultimately, the best way to learn how to work with AI is by actually using it in daily work.
The Decoder: Prompt engineering is becoming an important future skill. How do you assess its relevance for administration?
State Secretary Stich: Prompt engineering is truly a key competency for using AI effectively. That's why we need to integrate it into training programs. In Rhineland-Palatinate, for example, the Development Agency RLP has been conducting training specifically for local politicians and administrative staff for several years to make them proficient in prompting. It's a good start that we can build upon.
The Decoder: To what extent do you consider models like the "Gray Box Model of Chatbot Didactics" or an "AI Chatbot Learning Framework" suitable for training administrative staff?
State Secretary Stich: Such models are definitely helpful because they combine two important things: transparency and simple learning approaches. They make it easier for employees to understand how AI actually works, which builds trust. And when you understand how such a system operates, you can use it much more effectively.
The Decoder: What activities are planned regarding "citizen benefits through government AI use"?
State Secretary Stich: With AI, our main goal is to make administrative processes faster and simpler for citizens. In Rhineland-Palatinate, we already have initial examples, such as automated processing of housing benefit applications. In the future, we want to build an "AI Service Portal" that provides personalized information and support. In the short term, we're working to offer more AI-supported services. In the medium term, we aim to make these services accessible to ensure everyone can use them. Long-term, it would be great if we could fully automate and personalize many processes - making administration more modern and citizen-friendly.
The Decoder: How do you think administrative culture and structure need to change when government and citizens use AI tools for communication?
State Secretary Stich: I'm happy for all employees whose workflows are simplified by AI. It's a win-win situation when processes become faster and simpler while increasing speed and transparency for citizens. To get there, we need more openness, transparency, and enthusiasm for new things in some areas. At the same time, we need to implement agile working methods and continuously strengthen digital skills to handle the requirements of AI-supported communication effectively.
The Decoder: What changes do you see needed in training and development for state and local administration employees regarding AI?
State Secretary Stich: In the short term, it's mainly about teaching the basics and offering training, so employees know how to use AI. In the medium term, we need to integrate the topic firmly into basic and advanced training to make it accessible to everyone. Long-term, we need a true learning culture where employees stay fit to keep pace with constant technological developments.
The Decoder: How do you personally use AI chatbots like ChatGPT, and what do you consider most important about it?
State Secretary Stich: I use AI chatbots exclusively for private purposes and only when I have a question that can't be answered quickly and satisfactorily by a regular search engine. What's important to me is to critically examine the answers and understand AI as a tool that complements human thinking but doesn't replace it.
