Technical articles · 2025.09.12
Reinhard Stolle on vehicle ML uncertainty and lifecycle assurance
Interview mit Reinhard Stolle – »Die beste Technik auf sichere Art und Weise ins Fahrzeug bringen«
Overview
In an interview with Hans-Thomas Hengl, Reinhard Stolle discusses uncertainty in ML-based perception and limits of conventional software assurance. He connects AutoDevSafeOps and Safe AI Engineering research with lifecycle updates, V&V and monitoring, and discusses LLM support for engineering documents and foundation models for driving strategies as research possibilities.
Publication and version information
The source records posting on 2025-09-12 and modification on 2025-10-20; the update date is not an exact revision identifier. This is an interview, not product-validation evidence or a demonstration that vehicle LLMs are safe. Hengl is the writer and Stolle the interviewee. Standards names, vehicle approvals and project achievements are treated as source statements, without independently establishing current status or ISO requirements.