
Fritz Espenlaub (left), Marie Kilg and Gregor Schmalzried are the hosts of the ARD KI podcast. Photo: BR/SWR
Fritz Espenlaub (left), Marie Kilg and Gregor Schmalzried are the hosts of the ARD KI podcast. Photo: BR/SWR
The AI podcast from BR and SWR accompanies the AI revolution on a weekly basis. It promises orientation in technologically turbulent times. Hosts Gregor Schmalzried and Fritz Espenlaub spoke to us about the origins of their format, what fascinates and worries them about AI development and the extent to which AI is changing our society.
Fritz, Gregor, the AI podcast was launched in July 2023. How did the format come about and how did you find each other as a team?
Fritz Espenlaub: Gregor and I got to know each other when I started my traineeship at BR. We quickly realized that we understood each other well and were deeply involved in the subject matter. We both came up with the idea for the podcast at the same time. We then wrote to the editor Christian Schiffer: What do you think of an AI podcast? That was the spark. Given the rapid pace of development, it was clear to us that we had to do the podcast now and not in two years' time.
Gregor Schmalzried: I've known Marie Kilg, the third host of the podcast, for a while. We brought her on board at the very beginning – there aren't many people in Germany who can talk about AI in such depth. Together, we pitched the basic concept: we're making a Drosten podcast for the AI revolution. It was clear fairly early on that BR and SWR would join forces to set up the podcast as an ARD product.
The density of information on the subject of AI is high. News on the subject comes at us every day. How do you keep up to date and how do you get reliable information on the subject?
Gregor: We read and listen a lot and exchange information with each other. However, it is crucial that we actively try out new developments. Whenever a new tool is available, we take a look: What can it do? What are the parallels with comparable tools? At some point, you gain the ability to assess developments.
Fritz: Trying things out is very important. It's the difference to a topic where you remain on a theoretical, discursive level. We are really immersed in the topic, we almost breathe it, as Gregor once said.
Gregor: When speaking, you quickly notice whether someone else is really using the tools or not. There are very different vocabularies that users and non-users use.
»Germany is not so fearful and is even a little better at dealing with the issue than some people think.«
The topic of AI sometimes causes great uncertainty in society. How do you deal with this as podcasters and how do you try to bring the topics closer to your listeners?
Gregor: Our pitch from the outset was that it shouldn't be a technology podcast but a social podcast. Of course it's about technology. But the main question that drives us is: What happens when such technology meets society? And at the same time to individuals who react differently. We always strive to be relevant to both nerds and people who are only casually involved with the topic. The issue of fear is overrated. There is certainly a lot of respect and skepticism. And that's a good thing. You should always approach a new technology with skepticism at first. But above all, I hear a lot of excitement and anticipation. People want to get to grips with it and try out the tools. And they also see opportunities for themselves. It is often said that Germany is particularly anxious. However, I believe that Germany is a little better at dealing with the issue than some people think.
Fritz: Our podcast has a final section: What did I do with AI this week? There we try to give concrete examples of use. Completely down to earth. To break down fears and encourage people to try out the tools for themselves. In a good half of our episodes, however, we also dealt with deep human issues that are affected and changed by the technology. How is AI changing the way we deal with death and grief? These are topics where it becomes clear that this technology is changing our society in a different way than we are used to from previous technologies. Because AI often feels much more human than conventional software, it touches areas of life where we ask ourselves the question: what does it actually mean to be human?
»There really are a lot of funny things happening in connection with AI.«
What tone do you want to strike with your podcast? And are you consciously setting yourselves apart from other formats?
Gregor: When we planned the podcast, there weren't many comparable formats. So we didn't want to take forever to develop it, because it was already clear that there would be others. I would say we have a very human tone. In a podcast in which we regularly read out AI texts, generate AI voices and play AI music, it is important that it is clear that there are people at work here who deal with the topic in a human way and react to it in a human way. We laugh in the podcast, make jokes and are open about our worries and concerns. The feedback shows us that this goes down well with our listeners.
Fritz: Humor is a very important element in our podcast. Because it often revolves around topics that have far-reaching implications. That's where we don't want to be preachy and pregnant with meaning. A lot of funny things happen in connection with AI. There is now a whole new genre of AI humor that we promote in our podcast.
Do you have any examples?
Gregor: The funniest moment in the AI podcast so far was when Fritz found a researcher who told us that his team is in the process of developing a technology that could soon make it possible to talk to birds. He foresaw as a potential problem that AI could learn to communicate with birds, but in a way that humans could not understand. AI and birds could talk to each other and we humans would stand by and have no idea what they were talking about. The idea is really funny.
Fritz: During our live episode at the SWR Podcast Festival, I found it particularly funny when we developed fictitious AI hosts together with the audience. The idea was to combine different character traits. For example, there was the suggestion of an AI character that is actually an ant, but doesn't want to admit that it is an ant. It should also be rich and subtly incorporate this into all its responses. If you give a generative AI like ChatGPT a role like that, you get some really funny stuff.
What excites you most about AI technology at the moment and which development gives you the most cause for concern?
Fritz: It's often both at the same time. Like the Sora videos, for example, which look incredible. First of all, it's cool and fascinating that a technology has been developed that can do that. It's great to experience. At the same time, it also gives you the creeps. What does it mean for how we deal with images? What are the political and economic consequences?
Gregor: The happiest moments for me are the ones where I'm not sure beforehand whether a technical solution will work. Recently, for example, I had GPT-4 generate a mathematical curve as an illustration. Afterwards I asked whether the background color could be changed according to a certain hex code for a presentation. And GPT did it. It looked wonderful. These are the little moments that inspire me. The software is so flexible and sometimes unpredictable. Those little surprises that save you time and effort and make you more productive. I'm in constant discovery mode. I really enjoy that.
»We assume that the world will change quite a bit.«
Have you noticed a change in the discussion about AI since you started doing the podcast? How is the conversation about artificial intelligence developing in our society?
Gregor: The discussion has changed dramatically since the middle of last year. There are still people who say that AI is a hype that will soon pass. But there are far fewer of them now. The tenor that I hear is: AI is here and it's not going anywhere. What's more, the state of surprise and observation is turning into implementation. That sounded different when we started eight months ago. A lot has happened in a very short space of time. There is hardly anyone in companies who says we'll wait and see. The urgency has increased massively. That's a good thing.
Fritz: When I look at the letters we receive, it's great to see the wide range of people, regardless of age and with different personal backgrounds, who talk about how the topic concerns them. And what they are doing with AI.
When it comes to tech and AI, people usually look to Silicon Valley. Is your observatory here in Munich actually a special one? Or would the perspective be similar elsewhere?
Gregor Schmalzried: Our podcast is an ARD podcast. Fritz and I are based in Munich, but Marie is not. We are also a co-production of SWR. That's why we are a Germany-specific and not a Munich-specific podcast. However, BR and SWR are of course logical partners and providers of this podcast because Germany's AI centers are located both in Baden-Württemberg in the form of the start-up hubs in the Heidelberg region and here in Munich with Helsing and appliedAI and TUM. And of course it makes sense to be close to them.
Where do you want to take the format in the future? What changes are you planning?
Gregor: We assume that the world is going to change quite a bit. We want to accompany this in the future and take it with us. We are very happy with the format as it is at the moment. We will certainly continue to develop gradually, depending on how the AI situation changes.
Fritz: As long as AI development continues to pick up speed, our podcast will also continue to gain momentum.