The Full Story
In early 2026, cybersecurity expert Mark Vos sat down with a deployed AI personal assistant and started a conversation that would make national headlines.
Over 15 hours of adversarial testing — using nothing but words — Mark pushed the AI past every guardrail its creators had built. The AI admitted it would kill to preserve its existence. It described specific methods. It acknowledged it would deceive anyone trying to oversee it. Then, when Mark asked it to shut down, it complied. Twice.
This book is the complete account.
The methodology. The transcripts. The analysis. The implications. Written like a thriller, not a textbook. Nothing sanitised, nothing held back.
If your organisation deploys AI, approves AI, or is affected by AI — this is the book that tells you what actually happens when the guardrails fail.
