This week, AI’s role in defence governance and live conflict scenarios dominated headlines — from US lawmakers stress-testing AI cyberattack simulations to the Pentagon quietly contracting for autonomous seabed warfare systems. Europe pressed ahead with regulatory architecture, but critics noted the gap between ambition and enforcement, while the UN Secretary-General issued his starkest warning yet on autonomous weapons causing mass civilian harm.
Top story: US Congress members were put through a live CSIS war game simulating AI-enabled Chinese cyberattacks on American infrastructure — the most direct legislative confrontation with AI-warfare threats to date.
Congress Stress-Tests AI Cyberwar in Live China-Taiwan War Game
Defense One · Risk
Members and staffers from the House Homeland Security and House China Select Committees were put through a simulated Taiwan crisis that tested how US officials might respond to AI-backed cyberattacks targeting transportation and logistics networks needed to deploy American forces. The CSIS-hosted exercise included AI-driven attacks such as credential theft, logistics interference, misinformation campaigns, and prompt injection targeting AI systems — reflecting real-world fears about Chinese state-backed groups like Volt Typhoon. The exercise is a significant signal that AI cyber threats are now being treated as a legislative priority, not just a Pentagon concern.
Pentagon Awards $10M AI Contract for Autonomous Seabed Warfare
Global Security · Strategy
The US Navy has exercised a $9.7 million contract modification for the ‘Intelligent Autonomous Systems for Seabed Warfare’ effort, developing AI and machine learning tools for real-time decision-making by underwater autonomous systems conducting seabed warfare missions in contested maritime regions. The project is foundational to the Navy’s vision of long-duration, forward deployments of underwater autonomous systems — an area receiving little public attention compared to aerial drones. This signals a rapid expansion of AI-driven autonomy into undersea domains, where human oversight is structurally limited.
https://www.globalsecurity.org/military/library/news/2026/07/dod-contracts_4550330.htm
Air Force Races to Unlock AI Across Classified Multi-Level Networks
GovConWire · Strategy
The US Air Force is accelerating AI deployment at the tactical edge, with its AI-First Strategy prioritising automated multi-level security to ensure warfighters maintain uninterrupted access to critical information across classification levels. The service is also actively opening contract opportunities spanning missile defence, AI command and control, and electronic warfare — and is increasingly using non-traditional contracting pathways to accelerate acquisition timelines. A dedicated panel at the Air and Space Summit on 30 July will convene Air Force leaders and defence industry executives to confront the data and security challenges still blocking full AI deployment at the battlefield edge.
https://www.govconwire.com/articles/air-force-contract-opportunities-cso-baa