OT, ICS & SCADA field guides · Industrial AI

Industrial AI Security All-in-One by Kai London

In an office, a wrong model output is an embarrassment. On a plant floor it can move something heavy. This handbook covers AI where physical consequence is on the table — optimisation, digital twins, vision systems, robotics and the agents now being given a say in how the process runs.

Available on Amazon.

What is inside

Where AI sits in the plant

Advisory, supervisory or closed-loop — and why that distinction determines every control that follows.

Digital twins and simulation

Protecting the model of the process, and the risk of a twin that has quietly drifted from reality.

Machine learning on process data

Data integrity, sensor spoofing and drift in predictive maintenance and quality models.

Robotics and autonomous operations

Safety interlocks, physical boundaries and the relationship between functional safety and cyber security.

IIoT and edge devices

Provisioning, identity, update paths and the sheer device count that makes manual management impossible.

Agents with plant privileges

Approval gates, bounded actions and human override for AI permitted to influence the process.

Who it is for: manufacturing and OT technology leaders, automation and controls engineers, industrial data science teams, and CISOs approving AI in operational environments.

IEC 62443IEC 61508NIST AI RMFISO/IEC 42001NIST SP 800-82

Safety does not delegate. An AI may advise the process; it does not get to overrule the interlock.

About the author

Professor Kai London — CISSP, CISM.

An internationally recognised cybersecurity executive, board advisor and Founder & CEO of Quantum AI Systems Security LLC, writing at the convergence of AI, governance and operational resilience. Honorary Professor and Researcher at UCL.