DR PETER SPAYNE

HOW DO YOU TRUST

A KILLER ROBOT?

How to Trust a Killer Robot is the title of Dr Peter Spayne's doctoral thesis, completed at the Cranfield Centre for Defence Engineering (viva passed July 2026). It asks a question that sounds simple and is not: how do you assure a weapon system when the machine is both the equipment and the operator?

​

A lethal autonomous weapon is not a robot making one decision. It is a socio-technical system that distributes risk across designers, commanders, operators, regulators and the machine itself. Trust in such a system cannot rest on the confidence anyone has in a single component; it has to be built into the argument that connects all of them.

​

For crewed systems, engineers certify equipment as safe to operate, and organisations train people to operate it safely. Autonomy collapses that distinction. Peter's research merges the two concepts into a single assurance methodology, Safe to Operate Itself Safely, for weapon systems containing artificial intelligence. It is published in Defence Studies and was recognised with the IMechE Patron's Prize at INEC24.

​

The methodology reframes assurance from asking whether the machine is safe to asking whether the whole system can be shown to be safe, and whether someone can stand behind that claim. It gives defence organisations a way to build safety cases, assurance matrices and evidence chains for autonomy that hold up to scrutiny, rather than trusting a demonstration and hoping.

​

The peer-reviewed papers, conference proceedings and House of Lords written evidence behind the thesis are collected on the Research page.