What you will find

Three kinds of writing. Field notes on what we see across real AI estates, including the failure modes that keep recurring. Technical breakdowns of specific attacks and defences, with enough detail to be useful rather than merely alarming. And practical governance: how to produce evidence that survives contact with an auditor.

What you will not find

  • Vendor comparison pieces dressed up as research.
  • Speculation about capabilities nobody has shipped.
  • Client names or anything traceable to a specific engagement.

Topics we cover

AI Governance7 Threat Research5 Engineering4 Model Security3 Compliance1

8 articles published so far. You can follow along by RSS, or on LinkedIn.

Test your own team

We also run a free skills assessment covering secure AI practice, from foundations through to adversarial machine learning. Six courses, scored instantly, with a certificate on a pass.

Go to Certify