15.04.2026
Michael A. Hedderich Featured in “KI Pro” Podcast
Discussing Trust, Limitations, and Evaluation of LLMs From an Industry Perspective
MCML Junior Research Group Leader Michael A. Hedderich was recently interviewed by heise for their industry-focused “KI Pro” podcast. In the episode, he discusses key questions around large language models (LLMs), including how we can better understand their limitations and improve trust in their outputs.
The conversation offered valuable insights into how these topics are perceived from an industry perspective, particularly among practitioners and small and medium-sized enterprises (SMEs). It also highlighted the growing importance of reliable evaluation methods for LLMs in real-world applications.
The discussion builds on recent research conducted at MCML, including a Spotlight project on new approaches to evaluating LLM outputs, as well as a study on the reliability of knowledge probes for LLMs led by Raoyuan Zhang.
The podcast episode is available in German via heise’s “KI Pro” format.
Related
12.11.2026
One Toolbox for Many Kinds of Uncertainty
MCML researchers introduce probly, a unified Python package for uncertainty-aware machine learning, presented orally at NeurIPS 2026.
10.11.2026
Enkelejda Kasneci: The Real Value of AI in Education Is More Time for Humans
Enkelejda Kasneci explores how AI can provide personalized learning support while keeping human interaction and student agency at the center.
09.11.2026
Outstanding Paper Award at EMNLP 2026 for Work on Accent Bias in Audio Language Models
A paper co-authored by MCML PI Valentin Hofmann on accent bias in audio language models won an Outstanding Paper Award at EMNLP 2026.