Visionary of a citizen-centred, sovereign and future-proof administration.

Mr Beck, you believe we can only stop demographic change if we rethink processes. How can we go about that?

Demographic changes are going to have a huge impact on public administration. In the next ten years, we are set to lose around a quarter of the staff and over the next twenty, that will shoot up to 50 per cent. To counteract this impending crisis, we are going to have to rely more on automation and new technologies and artificial intelligence (AI) will play a key role in streamlining processes. It’s important to manage the transformation from the get-go to establish sustainable solutions and ensure our public authorities are able to carry on delivering in the long term.

What role is the Baden-Württemberg Innovation Lab playing in all this?

The InnoLab_bw is opening the door to new opportunities in public administration. Free from the constraints of classic structures, we can push on with innovating. In most cases, public authorities are stymied by prescribed processes, but our idea was to create a unit that sits outside of the hierarchical organisation—similar to how a company operates. This involved talking to start-ups, developing new solutions and trying out new technologies. Interestingly, we discovered that the corporate and public sector worlds are facing pretty similar challenges, for example, when it comes to interfaces, outdated IT systems and overly complex structures.

What about protecting citizens’ personal data?

Data protection is critical, particularly when it comes to using AI. Here in Baden-Württemberg, we attach a great deal of importance to not simply handing out our citizens’ data to external service providers. Whereas individuals can decide for themselves if they want to save information in the cloud, they don’t have the same control over what the state does, which is why we’ve put our faith in a sovereign infrastructure, run a lot of AI applications in-house and utilise the Baden-Württemberg State IT Authority’s (BITBW) data centre to ensure secure IT operation.

Let’s talk about the people. How do you convince employees about the benefits of these new digital applications?

New tools will be accepted if they bring real added value. If an application is just a digital carbon copy of an old process without any tangible optimisations, it won’t be used. When electronic document management was rolled out, the result was less time-consuming work such as manually sorting through and sharing files. While it did create some new work, overall many processes have been streamlined. Palpable benefits like these are exactly what are needed to get employees on board with new digital solutions.

So it’s primarily about adding value?

Yes, it’s the decisive factor. I can give you another example, this time from the judiciary. Family court judges often have to calculate maintenance payments and pension adjustments, but they can use specialist tools that considerably speed up the process. You won’t find any judge who would be happy to try and work it out on a piece of paper when they can just use a tool that will give them an accurate calculation in double-quick time.

You developed a tool like this with Aleph Alpha. What is F13 and how can it support public-sector employees? 

F13 is a text assistant based on generative AI that takes on routine tasks. It offers chat features, can summarise texts, make notes and help with research. We wanted to create a secure alternative to the publicly-available AI tools out there, which brings us back to data protection. Public authorities cannot enter confidential data into public AI tools, which is why F13 processes everything in the state’s own data centre—a secure environment tailored to the needs of public administration.

And what are the next steps for this tool? 

We are also making the technology available to other government organisations in the hope that as many German federal states as possible will start using this standardised software. When we started out with F13, we wanted to make it scalable, and asked ourselves how we could make a tool that works for 10 or 100 people work for hundreds of thousands or even millions. We now have hundreds of public authorities asking about the tool, which is quite the success. But we aren’t resting on our laurels. The team is already working on the next innovative and unconventional ideas.

What other developments do you see on the horizon?

I think we are on the cusp of a new era. AI models are still pretty inefficient and prone to errors, but there have been some very promising developments towards smaller, more powerful models. One big trend is going to be the move towards agent models. Right now, we enter single prompts, but in the future, AI agents could take on tasks themselves. For example, one agent would check my calendar, while another automatically schedules meetings and yet another handles red tape. This means that the technology will be taking on a greater number of admin tasks, which allow public-sector employees to focus more on their citizens. And when we think back to the imminent staffing crisis I mentioned before, it’s these kind of developments that we need to actively cultivate.