Newshot

Germany Raises Cyber Threat Assessment to 'Severe to Critical' as AI Agents Lower the Bar for Attacks

AI summary

Germany's Federal Office for Information Security (BSI) has raised the threat level facing IT systems to "severe to critical," noting that agentic AI systems are giving attackers more opportunities. In a report released on the 19th, the agency said many companies and consumers believe they are safe and underestimate the risks they actually face, while attacks are becoming faster, cheaper and more targeted, with its cyber security index jumping from 124 points in mid-2025 to 238 points in mid-this year. Recent disruptions to supermarket refrigeration systems in Europe, French blood banks and German hospitals have highlighted the threat to the interconnected systems underpinning the modern economy, and high-performance AI models can now find and exploit vulnerabilities faster and across a wider scope.

Translation(model translation, not the source's own text)

Germany Raises Cyber Threat Assessment to "Serious to Critical": AI Agents Lower the Bar for Attacks

IT之家 November 19 news, according to a report by Reuters this afternoon (the 19th), the German Federal Office for Information Security (BSI) has raised the threat to information technology systems to "serious to critical." The office noted that agentic AI systems are giving attackers more opportunities.

In a report released on the 19th local time, the office said many businesses and consumers believe they are safe and underestimate the risks they actually face. The assessment is published annually; last year the office judged the situation as "serious."

In recent months, refrigeration systems at supermarkets in several parts of Europe, France's blood banks, and hospitals in Germany have all experienced operational problems due to cyberattacks. These incidents highlight the threat facing the interconnected systems that underpin the modern economy.

Another high-profile cybersecurity incident in recent weeks involves Java, a commonly used programming language. On Wednesday, the German Federal Office for Information Security and governments of other countries issued warnings that a serious vulnerability exists in Java that hackers could exploit to gain access to computer systems.

The report states that in the past, exploiting a vulnerability required specialized knowledge and took some time. Now, attacks are becoming faster, cheaper, and more targeted.

The German Federal Office for Information Security measures the level of threat using a cybersecurity index. The index rose sharply from 124 points in mid-2025 to 238 points in the middle of this year.

The report notes that high-performance AI models can discover and exploit vulnerabilities faster and across a wider scope, shifting the threat landscape during this assessment period toward "critical."

Attackers have begun using AI models to directly attack systems, with email-based attacks declining as a result.

As Germany's digitalization deepens and connected devices become more widespread, the number of entry points available for intrusion and the weak points that could be exploited have increased considerably, expanding the overall "attack surface."

Governments have made progress in tightening regulation of information technology security, but the report warns that many businesses are too slow to strengthen their defenses, especially when it comes to using AI systems to scan for vulnerabilities before an attack occurs.

The report states that small and medium-sized enterprises sustain the German economy but often have few means to defend against attacks. Operators of critical infrastructure are subject to stricter regulation and generally have better protections, but there is still room for improvement.

Related Articles

  • Cyber threats intensify in the UK financial sector, with the number of incidents reported to regulators in October hitting the second-highest single-month level in five years
  • Anduril founder Luck on AI cybersecurity: The US faces an AI offense-defense imbalance, and in the long run the side that makes good use of AI for defense will hold the advantage
  • Multiple well-known websites defaced, attackers steal cryptocurrency and breach iPhones
  • Marsh: There are signs that hackers are using AI tools and agents to find security vulnerabilities
  • Best birthday gift: Fan fiction platform AO3 comes back online on the same day as its 17th anniversary
  • Beware of scams touting a PC version of "GTA 6": They can steal Steam accounts and crypto wallet data

The translation is machine-generated and individual terms may be off. Nothing is rewritten or cut; the original is authoritative.

Published by IT之家
This site only summarises and aggregates; copyright stays with the original author. Headlines and summaries follow the source — where they differ, the original wins.
Read the original ↗