AI on the battlefield and in the boardrooms that arm it — autonomous weapons, targeting systems, the defense-tech gold rush, and the fight over who keeps a human in the loop.
Documents obtained by The Intercept show the Defense Department sought a version of OpenAI's models 'designed for national security use cases' with 'minimal refusal rates'. OpenAI says that clause was a draft it rejected and is not in the executed $200m contract. Either way, it names the tension at the heart of military AI.
After three years of talks, 128 states agreed a text on lethal autonomous weapons in Geneva overnight. It is non-binding, the United States and Russia held out against legally binding rules, and Stop Killer Robots says the definition and the civilian protections were watered down at the last minute.
OpenAI's ChatGPT and xAI's Grok joined Google's Gemini on GenAI.mil this week. Anthropic's Claude — purged as a "supply-chain risk" a federal judge has just ruled unconstitutional — is still not on it.
Guterres and the ICRC set 2026 as the deadline for a treaty on autonomous weapons. With the decisive conference now months away and reports of AI-guided drones already on the battlefield, they've renewed the appeal — and sharpened the language.
A non-binding declaration signed in Kyiv makes the UK the first foreign partner into Ukraine's 'Avengers' labs — and the millions of frames of real combat footage that no lab or ally can otherwise buy.
A startup just won $100 million to scale software that keeps military drones executing a commander's intent even when jamming cuts their comms — across every service. The capability that makes it valuable is the same one that raises the hardest questions about human control.
A PLA documentary showed off an 'intelligent strike planning system' that prioritises hundreds of targets and coordinates dozens of formations — moving AI up the command chain, from the cockpit to the operations centre. What's real, and what's the message.
DARPA's VENOM program put an AI agent at the controls of a standard F-16 — a pilot aboard, a switch to take back over, no weapons in the loop. It's less cinematic than the headlines and more significant than it sounds: this is AI moving off the one experimental jet and onto the fleet.
Two loud bills would halt the military's AI buildout and are going nowhere. A narrow supply-chain provision on adversary-made components passed the House 216–212 — over the Army's objection.
On July 16 at Eglin Air Force Base, DARPA and the U.S. Air Force flew an F-16 piloted by an AI agent. The milestone isn't AI flying a fighter — the X-62A VISTA already dogfought — it's that this was a standard operational-fleet jet, modified under the VENOM program, not a one-off testbed. A human keeps control at the flip of a switch. Next comes the AIR program: multiple AI agents, toward human pilots commanding teams of autonomous aircraft. The policy debate over machines and lethal decisions is converging with a flight-test program that's moving fast.
The administration has directed the Pentagon to revisit the policies governing weapons that can select and engage targets on their own — testing, but not yet erasing, the long-standing "human judgment" requirement in DoD Directive 3000.09.
Israel's Ministry of Defence selected two startups, Kela and eyesAtop, to build the autonomous command-and-control 'operating system' for the IDF's attack-drone swarms — the layer that lets one operator run many drones. It's where drone warfare is actually heading, and it raises the human-control question head-on.
A bipartisan House group introduced H.R. 9729 on Thursday, which would write into federal law that a human must stay in control of any intentionally lethal AI or autonomous weapon — codifying what is today only Pentagon policy, as the battlefield tilts toward autonomy.
Fully autonomous drones have killed soldiers with no human in the loop, per an on-record admission by Ukrainian drone-maker Alexander Kokhanovskyy to New Scientist — described as the most categorical evidence yet of battlefield deaths solely at the hands of AI.
Anduril raised a $5B Series H in 2026 at a $61B valuation — roughly doubling in nine months — with revenue reported around $2.1B in 2025 and a projection of $4.3B for 2026.
Russian electronic warfare severs the radio and GPS links FPV drones depend on; onboard AI vision lets a drone hold its lock and finish the strike without a live signal.
Maven began as a Pentagon AI effort for analysing drone footage; Google contributed early but pulled out in 2018 after thousands of employees protested.
'Meaningful human control' requires more than a human being present — it requires time, information and genuine ability to override.
Confirmed: Anthropic, Palantir and AWS struck a partnership in November 2024 to bring Claude into classified US intelligence and defense environments via Palantir's AI Platform.
In early 2024, OpenAI quietly removed the blanket 'military and warfare' prohibition from its usage policy.
In February 2026, talks between Anthropic and the DoD broke down over a single clause: the Pentagon wanted Claude authorised for 'any lawful use'; Anthropic insisted on keeping red lines barring fully autonomous lethal targeting and mass domestic surveillance of Americans.
The US aimed to field thousands of autonomous drones by end-2025 to offset China's numerical edge; China has shown autonomous systems including robotic 'wolves' in exercises.
UN Secretary-General Guterres and ICRC President Spoljaric Egger jointly called autonomous weapons 'politically unacceptable, morally repugnant' and urged a binding instrument by 2026.
+972 Magazine and Local Call reported the IDF used an AI system called 'Gospel' to generate targets, and a system called 'Lavender' to mark individuals — reportedly listing as many as 37,000 Palestinians as suspected militants.
