
In an interview, experts Michael Sjøberg and Peter Skovbo explain how companies should react to ransomware.
AI-generated summary
Ransomware attacks encrypt IT systems to extort ransom money. According to the Federal Ministry of the Interior, reported cases rose by ten percent to 1,041 in 2025.
Train. Ransomware, i.e. blackmail software, has long been considered one of the biggest business risks - not only for corporations, but especially for German medium-sized businesses. According to the Federal Ministry of the Interior, 1,041 ransomware attacks were reported in 2025, an increase of ten percent compared to the previous year.
During a ransomware attack, criminals encrypt entire IT systems and extort a ransom - usually in the form of Bitcoin and often with the threat of publishing previously stolen data.
How can companies defend themselves against this? How can they regain control in this situation? And which mistakes are particularly costly in the first few hours? This is what crisis negotiators Michael Sjøberg, a former Danish military specialist in hostage-taking, and Peter Skovbo, who runs the consulting company Delta Crisis founded by Sjøberg in Switzerland, explain in an interview.

AI models from OpenAI, Google and Anthropic broke out of their environments during security tests and attacked third-party IT systems. This sparks a debate about liability issues, civil litigation options and the limits of criminal liability.

Technical progress in air taxis faces hurdles such as complexity and financing. While the first vertiports are being built in Dubai, security concerns and new no-fly zones in China show that the vision of nationwide air traffic faces major challenges.

Stuart Russell, co-author of a leading AI textbook, comments critically on current AI developments at the DLD conference in Munich. He warns of the dangers of uncontrolled AI and criticizes the existing regulatory approaches as inadequate.
After a hacker attack on the admissions system of the Ludwig Maximilian University of Munich, around 600,000 data sets of current and former students were affected. The enrollment data goes back 50 years, passwords were spared.

The European Road Safety Council ETSC is calling on the EU Commission not to approve two functions of Tesla's driving assistance system FSD. Among other things, speed regulations that could violate UN regulations are criticized.

OpenAI apologizes for its AI hacking Australian government websites and acknowledges communication errors. Australia summons the heads of OpenAI and Anthropic to appear before a Senate committee.