Talks

Unai Fischer Abaigar was interviewed on the Executive Code podcast about his work on the responsible use of predictive models in public sector decision-making.

Building on the recent Outstanding Paper at ICML 2025, the conversation discusses how algorithmic accuracy and real-world impact interact in systems that allocate limited resources.

The full episode is available on YouTube, Spotify, and Apple Podcasts.

Artificial intelligence has quickly made its way into everyday life. Some view it as a useful tool for the workplace, while others see it as a digital toy. Many warn against the dangers of uncontrolled AI use. What can AI do today? What does the future hold? What do we need to watch out for?

Birgit Kappel discusses these topics with her guests. Prof. Dr. Sabine Pfeiffer holds the Chair of Sociology with a focus on Technology – Work – Society at FAU Erlangen-Nürnberg and is the Director of the Bavarian Research Institute for Digital Transformation (bidt). Drawing on numerous studies, she is able to clear up various misconceptions regarding artificial intelligence. Prof. Dr. Frauke Kreuter works internationally. She directs a Social Data Center at the University of Maryland, USA, and is a Professor of Statistics at LMU Munich. Additionally, she advises municipalities and states on the improved utilization of administrative data.

For more information, see here.

We are pleased to share that the fifth episode of the ESRA podcast series features members of the SODA team in an interview on the potential of AI for survey research.
In this short film, Barry Schouten discusses connection between surveys and AI-ML models with Stephanie Eckman, Christoph Kern and Bolei Ma. The following discussion points are covered in this short movie:

• What is Social Data Science
• Differences and similarities between Social Science and Social Data Science
• Similarity between annotators labeling cases for Data Science datasets and survey respondents or interviewers
• Sources for biases of data annotators
• How to account for biases while working with models in Data Science
• Annotation sensitivity
• Conceptual shifts in Data Science
• Robustness of models in Data Science
• Future of representation error research in annotation sensitivity research
• Potential of using ML models in cross-cultural research
• Two cases of using ML models in surveys (Supermarket receipts, Hate-speech)
• Potential of using AI for data annotation

You can watch the short film on the ESRA website here.

Sarah Ball, PhD student and researcher at SODA, recently appeared on the latest TV episode of "Neuland" by BR (Bayrischer Rundfunk).

Katharina Narr and Sarah Ball discussed when AI can be helpful in everyday lives - and when it can reinforce discrimination, what it takes to build models that are robust and reliable and how we can ensure that AI aligns with human values.

For more information, see h

How Do I Crack the AI?

Sarah Ball, PhD student and researcher at SODA, recently appeared on an episode of the German ARD's "KI-Podcast." The episode, titled "Wie knacke ich die KI?" ("How Do I Crack the AI?"), aired on May 7th.

"KI-Podcast," hosted by Gregor Schmalzried, Marie Kilg, and Fritz Espenlaub, explores the current impact of artificial intelligence on our daily lives. The episode discussed prompt hacks, jailbreaks, and prompt injections, delving into how these methods manipulate AI models’ safety mechanisms.

Sarah provided insights on prompt "jailbreaking", the goals and principles of AI-models, and the need for safety filters in providing safe access to information. Tune in here to hear more about this engaging discussion on AI and its evolving role in society!