This week’s defence AI stories share a single throughline: the race to operationalise AI at scale — across allied command networks, military platforms, and battlefield systems — is intensifying faster than governance frameworks can keep pace. From the Five Eyes binding themselves to a shared AI backbone, to NATO allies openly debating US tech dependency at Ankara, the central tension is no longer ‘should militaries use AI?’ but ‘who controls it, and under whose rules?’
Top story: New Zealand’s formal joining of Project Arcadia completes the Five Eyes AI battlefield command network, binding the UK, US, Canada, Australia, and New Zealand to a shared AI-driven operating picture — a historic and legally contested milestone.
Five Eyes Completes AI Battlefield Network With New Zealand Sign-On
RNZ (Radio New Zealand) · Strategy
New Zealand has become the final member of the Five Eyes alliance to formally join Project Arcadia, a US-led initiative designed to build a shared, AI-enabled battlefield command network capable of matching the attack speeds seen in recent conflicts. The project creates a unified Common Operating Picture across all five nations simultaneously, with the UK’s Digital Targeting Web feeding directly into the system. Critics, including New Zealand academics, have questioned whether smaller allies are ceding digital sovereignty to a US-first agenda, raising unresolved legal questions about accountability when AI-assisted decisions lead to catastrophic outcomes.
https://www.rnz.co.nz/news/world/674778/new-zealand-joins-us-project-to-accelerate-ai-warfighting
NATO Ankara Summit: AI Tech Supremacy Splits Alliance Along Transatlantic Lines
IBTimes UK · Regulation
As NATO leaders gathered in Ankara on 7–8 July, AI emerged as one of the summit’s most consequential undercurrents: the US arrived holding a decisive advantage through frontier models from OpenAI and Anthropic, while European governments expressed growing alarm at their dependence on technology subject to Washington’s unilateral decisions. The Trump administration’s tightened access controls on advanced AI systems have accelerated Europe’s push for indigenous capability, centred on the Helsing–Mistral partnership developing military AI systems already tested in Ukraine. The dynamic is reshaping the balance of power within NATO itself, with AI access becoming a new form of strategic leverage.
https://www.ibtimes.co.uk/nato-summit-ai-defence-strategies-1807027
Grok Goes to War: Pentagon Deploys xAI Across 3 Million Military Personnel
eWeek · Generative AI
The US Department of War has expanded its GenAI.mil platform by integrating xAI’s Grok family of models, giving all 3 million military and civilian personnel access to frontier AI at Impact Level 5 — meaning it can handle sensitive Controlled Unclassified Information in daily workflows. A notable feature is that users will also gain real-time intelligence from the X (formerly Twitter) platform, framed by the DoW as a ‘decisive information advantage.’ The deal has drawn Congressional scrutiny over Grok’s history of generating unreliable content, highlighting the tension between deployment speed and responsible AI safeguards in a military context.
UK Army Reservists Build AI Tools for Frontline Policing and Readiness
The Defense Post · Tools
US Army Reservists have been developing and deploying AI tools spanning policing, logistics, and operational readiness — a signal that AI-enabled capabilities are moving beyond specialist units into the broader military workforce. The development reflects a wider trend noted by the Pentagon’s AI Acceleration Strategy, which demands ‘deployment velocity’ and real-world feedback loops measured in days, not years. For UK and European defence establishments watching closely, this bottom-up AI development model presents both an opportunity and a challenge: harnessing reservist ingenuity while maintaining appropriate oversight and standardisation.
FY2027 NDAA Mandates New AI Security Rules for All Defence Contractors
GovInfoSecurity · Regulation
Both the House and Senate versions of the US FY2027 National Defence Authorisation Act — authorising $1.15 trillion for defence — contain significant new AI security obligations for Pentagon contractors, including firmer CMMC cybersecurity expectations and fresh guardrails on how AI can be deployed across military networks. Pentagon cyber spending has grown to roughly $15.1 billion for FY2026, even as the civilian cybersecurity agency CISA faces severe budget cuts, creating a stark two-tier approach to national cyber resilience. UK and European defence contractors working with the US supply chain should prepare for these compliance requirements, which could reshape procurement standards across the Atlantic.
https://www.govinfosecurity.com/ndaa-drafts-bolster-contractor-cyber-ai-rules-a-32029