Artificial intelligence

28.05.2026

When FMEA meets AI
Why isolated silos create audit risks

What exactly is the failure mode that your FMEA is supposed to describe if a neural network detects pedestrians? Explain why traditional safety methods reach their limits when applied to in-vehicle AI.

21.05.2026

As the Operational Design Domain expands, traditional tests fail
SOTIF in automotive practice

An AI system passes all lab tests but still fails in the field. No bug, no programming error: ISO 21448 has a name for this exact phenomenon. What this means for your safety case.

04.05.2026

AI in software development
Why tools alone are not enough

AI tools have long been adopted by many development teams, yet the expected increase in productivity often fails to materialize. What’s missing between code suggestions and automated workflows, and how can this gap be closed?

22.04.2026

AI in Action
A Look Back at the Afterwork Exchange

How do you establish an AI culture within a company? What does the EU AI Act mean for regulated industries? At the first Afterwork Exchange, "Business Impact: AI,” hosted by sepp.med, decision-makers and experts discussed these questions. The answers were surprisingly concrete.

24.03.2026

Reasoning AI under scrutiny
How to make explainability verifiable

LLMs that use reasoning explain their answers in natural language. This opens up new possibilities for audits in regulated industries. One prerequisite is that the explanations themselves become part of the quality assurance process.

05.03.2026

The EU AI Act by industry
What distinguishes the medtech, automotive, finance, and administrative sectors?

The EU AI Act affects almost every industry, but not all industries are affected in the same way. What specific obligations apply to the medical technology, automotive, finance, and public administration sectors? Here’s a compact industry comparison.

27.02.2026

The EU AI Act in 2026
What every company using AI software needs to know now

The EU AI Act is the world’s first comprehensive AI law, and it impacts many more companies than most people realize. Here’s an overview of risk levels, deadlines, and affected industries.

18.02.2026

AI in the CI/CD pipeline
A hackathon demonstrates how it works

CI/CD pipelines generate thousands of log lines every day. At an AI hackathon, a team led by our colleague Yu Zou developed a tool that automates error analysis, solution proposals, and risk assessment. A field report.

16.02.2026

How to secure AI systems: Why traditional testing is not enough

AI software behaves fundamentally differently from traditional applications. Those who test it using conventional methods risk blind spots in audits and approvals. Which approaches will make a difference in 2026?