- Authors Georgia Cole and Isabella Wilkinson say the UK’s Ukraine AI labs access and U.S. autonomy partnership raise dependency and kill-switch risks.
- They warn that alleged AI targeting errors, escalatory behaviour in wargame studies and rubber-stamp human review raise risks, including nuclear escalation.
- They urge mandatory testing, transparency, clearer red lines, diversified partners, and a UK rethink on autonomous weapons treaty talks in November.
The United Kingdom’s Military AI Ambitions Need Guardrails

Loading comments…

