BSI AI Guideline

ByStefan Karg

What a good start into the weekend 😎

📄 The German BSI has published a guideline for developers regarding potential vulnerabilies in AI systems.

📣 From my point of view, a security risk assessment on systems using AI technologies has to consider the attacks on the AI parts of the system as well. A security risk assessment only considering the IT security aspects of such systems but neglecting the AI security of the system cannot be considered as complete any more.

You can find the pdf at this link when searching for “AI security concerns in a nutshell – Practical AI-Security guide”.

Similar Posts

  • Realtime AI applications

    ByStefan Karg

    Real time AI applications can bring huge benefits but also come with (social) risks we will have to deal with. The example below is a good demonstration of what is possible today. I’m very curious regarding the public discussion on how we will balance benefits vs. risks of such applications without killing technological progress in…

  • ISO 42001 Practitioner certification

    ByStefan Karg

    This month I’ve earned the ISO/IEC 42001:2023. Artificial intelligence management system practitioner certification of rigcert.education. As AI systems become widely integrated into the business world, securely managing them is increasingly critical. How can organizations ensure the reliability of AI outputs, protect company data from loss, and maintain system availability? These topics—and many others—are addressed by…

  • WormGPT and FraudGPT

    ByStefan Karg

    As with any new technology, the “other side” is trying to figure out how to use it to their advantage. The normal commercial versions of LLMs have barriers built in to prevent the creation of malware code or phishing mails, for example. If you remove these barriers, you get the output that we as security…

  • AI in the rail sector

    ByStefan Karg

    We are already seeing real-world applications of artificial intelligence in the rail sector. Deutsche Bahn is using artificial intelligence to reduce delays on its network. My prediction: This is just the beginning of ML/AI applications in our sector. We will see machine learning and AI algorithms becoming more and more widespread in the coming years….