AI red teaming vs penetration testing: what is the difference?
An AI application can pass a penetration test and still be manipulated through perfectly normal use. That is why red teaming has become the reference security test for AI systems
We use this place to discuss random subjects. You may find some of them interesting...
An AI application can pass a penetration test and still be manipulated through perfectly normal use. That is why red teaming has become the reference security test for AI systems
AI systems can be attacked in ways conventional cybersecurity was never designed to catch. ISO/IEC 27090, published in 2026, is the international guidance for addressing them — here is what it covers and who should read it.
Provider, producer or user? The role an organization plays in the AI ecosystem decides what its ISO/IEC 42001 management system must cover. An auditor’s guide to the roles defined in ISO/IEC 22989, with examples.
AI bias rarely comes from anyone intending it. Learn where it actually comes from — in the data and the design — and why it's so hard to spot.
Not sure whether to go for the ISO/IEC 42001 Practitioner or Auditor certification? They're not the same — here's how to choose.
How is artificial intelligence impacting the development and use of ISO standards
ISO/IEC 42001 and Microsoft Responsible AI Standard are two leading frameworks for responsible AI. They share a similar goal but differ in many ways.
AI has the potential to solve some of humanity’s greatest challenges. But we must use it wisely.
The AI system impact assessment is a key element of an AI management system according to ISO/IEC 42001. Why is it so important?
Thoughts on ISO/IEC 42001, the first international standard for artificial intelligence management systems