“In hostile conditions, a runtime guardrail should be rehearsed before an unlogged change makes it mandatory; the adversary already knows this.”

— Professor Kai London, CISO.

Quote card by Professor Kai London reading 'In hostile conditions, a runtime guardrail should be rehearsed before an unlogged change makes it mandatory; the adversary already knows this.'. From the book The AI Control Architecture.
Quote card by Professor Kai London reading 'In hostile conditions, a runtime guardrail should be rehearsed before an unlogged change makes it mandatory; the adversary already knows this.'. From the book The AI Control Architecture.

In hostile conditions, a runtime guardrail should be rehearsed before an unlogged change makes it mandatory; the adversary already knows this. — Kai London (Professor Kai London), CISO. Principle 2642 of 10000 from the book “The AI Control Architecture” — cybersecurity, AI security and OT resilience doctrine. Official sites: professorkailondon.com · kailondon.co.uk

Related: Principle 5142 Principle 7642 Principle 142

View in gallery (page 27) · About this book