Building an AI Security Function

The ambition to adopt AI is outrunning the ability to secure and govern it. This is a practical, principled guide to closing that gap: the foundations that endure, the operating model that works, and the roadmap that gets you there.

Understand the Gap

The gap between ambition and ability

Organisations and governments are racing to put AI into everything. Their ability to secure and govern it is not keeping pace, and in places it is going backwards.

The tools might work but what lags is the organisation around them: who owns AI risk, how fast assurance can move, and whether you have the right people to run the controls. That gap is where real risk accumulates.

See the Evidence

0.1%

of 2026 AI spending goes on securing AI: about $2.8bn against $2.52tn

Source: Gartner

68%

of breached organisations had no AI governance, up from 63% a year earlier

Source: IBM, 2026

82%

found AI agents they did not know existed in the past year

Source: CSA, 2026

42%

of UK security teams name AI as their most urgent skills gap

Source: ISC2, 2026

The Bletchley Park Model

In 1939, a diverse team of mathematicians, linguists, chess champions and crossword enthusiasts gathered at Bletchley Park to tackle an unprecedented challenge. They succeeded not through hierarchy or rigid doctrine, but through mission-driven collaboration, intellectual humility, and relentless ambition in service of humanity.

We propose this as the operating model for your AI Safety and Security function. The fundamentals haven't changed, but the stakes have never been higher.

Read More

Latest Articles

Practical guidance across every capability your AI Safety and Security function needs: from threat intelligence to governance, architecture to operations.