This week, Western governments moved from strategy to hard institutional action on defence AI: the UK published a mandatory directive embedding AI across its entire armed forces, the US Senate pushed to create a dedicated autonomous warfare combatant command, and a new White House executive order stood up an AI cybersecurity clearinghouse. The throughline is acceleration — the question is no longer whether to deploy AI in defence, but how fast and with what guardrails.
Top story: US Senate Armed Services Committee voted to create the first-ever Robotic and Autonomous Systems Combatant Command, signalling that AI-driven warfare is now a distinct military domain.
UK MOD Issues Mandatory Directive: Every Soldier Must Embrace AI Now
GOV.UK · Strategy
The UK Defence Secretary issued a new department-wide directive on 10 June 2026, making clear that AI is now a core part of how UK Defence must ‘deter, fight and win,’ with every part of the organisation expected to identify where AI can improve operational effect and speed up decision-making. Published alongside the launch of Taskforce RAID, the memo signals a fundamental shift from pilots and strategy documents to mandatory adoption at speed. Crucially, the taskforce has been granted exemptions from standard financial and procedural controls so it can move at the pace of the technology — a rare and significant concession in Whitehall.
US Senate Moves to Create Dedicated Autonomous Warfare Combatant Command
Defense One · Strategy
The Senate Armed Services Committee’s version of the FY2027 NDAA includes a provision to establish a new ‘Robotic and Autonomous Systems Combatant Command’ led by a four-star general — the first ever command dedicated entirely to AI-driven and drone warfare. Senators were inspired in part by Ukraine’s creation of a separate drone military service branch. The move reflects growing pressure to give autonomous warfare its own institutional home within the US military hierarchy, sitting alongside the Pentagon’s request for $55 billion for the Defense Autonomous Warfare Group in the same budget cycle.
OpenAI Publishes Pentagon Red Lines: No Surveillance, No Autonomous Weapons
OpenAI · Regulation
OpenAI has publicly detailed the specific guardrails in its classified Pentagon agreement, including hard prohibitions on using its models for mass domestic surveillance, directing autonomous weapons systems, or high-stakes automated decisions such as social credit scoring. The company says it retains full control of its safety stack and deploys via cloud only, with cleared OpenAI personnel kept in the loop — an architecture designed to make misuse structurally harder, not just contractually forbidden. The publication raises the bar for transparency in military AI contracting and puts direct pressure on competitors who accepted broader ‘all lawful use’ terms.
https://openai.com/index/our-agreement-with-the-department-of-war/
Trump Signs EO Creating AI Cybersecurity Clearinghouse for Critical Infrastructure
Holland & Knight · Regulation
A White House executive order signed on 2 June 2026 directs the NSA, CISA, and the Treasury Department to establish a new AI cybersecurity clearinghouse — a coordinating body to scan for software vulnerabilities, validate them, and prioritise patches across government and critical infrastructure within 30 days. The order also establishes a voluntary pre-release framework allowing AI developers to share frontier models with federal agencies for up to 30 days before public release for national security testing. Legal analysts note this is a significant reversal from the administration’s previously hands-off stance on AI governance, elevating the NSA into a central civilian oversight role.
UN Convenes First Global AI-Security Ethics Conference in Geneva
UNIDIR · Regulation
UNIDIR convened its Global Conference on AI, Security and Ethics 2026 at the Palais des Nations in Geneva on 18–19 June, bringing together diplomats, military officials, AI labs, and civil society to address the governance gap around AI in warfare. Sessions addressed agentic AI in cyber defence, autonomous decision-support systems, and whether AI-driven targeting analysis can genuinely reduce civilian harm under international humanitarian law. The conference follows UN General Assembly resolution 80/58 and marks a shift from debating AI’s military risks to seeking concrete implementation frameworks — a dynamic of direct relevance to UK and EU defence planners navigating their own governance obligations.
https://unidir.org/event/global-conference-on-ai-security-and-ethics-2026/