 Welcome to our Open Talk series. My name is Teri Zatzuga and I'm the research lead of the AI and Society Lab. I'm leading an interdisciplinary research group that wants to find out how AI can serve the public interest. And pretty early in our research we understood that this is such a big and complex question that we cannot solve it completely by ourselves. So in this series of conversations we speak to people who bring in their experiences and their research to find out the limits and the potential of AI to serve the public interest. And we really hope you enjoy these conversations. My name is Stefan Dreier. I'm one of the senior researchers, the one for media law and media governance at the Leibniz Institute for Media Research. And I've been working in this position for almost 20 years now. So I'm a lawyer by training and my research focuses on general transformation of media use, media power and the regulatory challenges that brings with it. Now I'm speaking to you as one of the PIs of the project financed by the Bavarian Institute for Digital Transformation. That project is called Coding Public Value. The main context and the problem that we want to clarify a bit is that in changing media landscapes there is ongoing pressure and public debate on how public service media companies or broadcasters can change with these developments. So one sentence regarding this public service media companies, those are the ones established by law in Germany and that have a specific public value remit. The idea here is that more and more media offerings in online environments and the question is if you use offerings in online environments you need software to do so and how these very general and broad remits can be applied to those software environments and especially to the engineering process of those software products. As you might already know this is not only a legal issue in the sense of regulatory requirements engineering but they are also deeply involved into public discourse and political discourse regarding the expectations that society has towards those broadcasters. The software engineering part plays a specific role within this project because we want to come up with ideas and options on how to translate those expectations into formalized software. How would you define public value? We try to come up with a definition of public value and public value is not even a vague legal term that you can find anywhere in legal texts. Public value is as itself an estimation or an expectation towards specific organizational structures and forms of conduct that might lead to public value as an outcome. You search for requirements necessary to the engineering process to achieve public value. Does this change the process of engineering generally? The answer is yes. This changes to a significant degree how you develop software because this is not only user or business model focused but you have to consider more public or stakeholder related requirements as well. This is kind of a deliberation. The deliberative part is already proceduralized in public service media. So they have internal boards, for instance, that check on how those concretization has taken place and whether this is still within the public value remit. There is no single true understanding of democracy. How do you translate this diversity into a proper media diversity? There's also a part of service offerings where I as a user might be able to decide on what algorithm I want to select for my content. And this is where we can apply the concept of diversity, not on the diversity of content but on the diversity of selection mechanisms that I can choose from. So this is very much a level of understanding of diversity that can now be applied to software of those public service media providers. So that's a very interesting thing to have because then we have to think about user interfaces, for instance. What does that mean? How can I explain to an end user that there are different understandings of democracy understanding based selection models for diversity or diverse content? What's your opinion on the challenge of translating requirements from the regulatory sphere into software development? We have discussions about this challenge almost every week between us and the software engineering colleagues. There's so much going away in translation already between lawyers and software engineers. If you look at the latest proposal on the AI regulation of the European Commission, it was I was shivering because there's so much implicitly going on into those provisions. So this is the first challenge coping with those legal terms that somehow should be applied or have to be applied into software without understanding from a legal perspective how software engineering works. And the second challenge that we saw is that legal understandings of words or norms change over time because of court rulings, for instance. And that is again not how software is being built. Once the system has gone live, you will optimize it towards a better user experience. But you are seldom using the software to get a feedback loop to the legal requirements. So we need a model that is also available in life systems to be sustainable or sustainably compliant with the legal requirements. And for that, you need someone who is also monitoring changes on the legal level. Our open talks, our open for collaboration, contact us to get involved.