How the systems themselves work, and what their design choices make possible.
Anthropic told courts it cannot alter Claude once it is inside a classified Pentagon system, and an appeals court has now ruled that how a laboratory trains its models can itself be a supply chain risk. Military AI splits responsibility between the laboratory that trains the model, the platform that gives it tools and the commander who acts, and no single party holds the whole of it.
Palantir says the Pentagon replaced Claude inside Maven with no operational impact. The claim highlights a different dependency: the platform connecting models to military data and workflows.
Ukraine's Starshield episode shows that a satellite terminal on a drone is a dependency someone else can adjust, and that Europe has no ready replacement. The harder question is operational resilience: which missions survive losing the link.
American campaigns have adopted generative AI faster than almost any tool in modern politics. Laboratory experiments show it can persuade attentive voters, while decades of field research suggest that campaign contact rarely produces measurable changes in candidate choice. The midterms will be read as a verdict on AI, though they were not designed to deliver one.
A pretraining researcher resigned from Anthropic this week and accused his employer and OpenAI of gambling with our lives. The part of his case that cannot be tested has taken the headlines. The part that can be tested is a set of cyber security incidents, and they describe a containment problem that is already here.
A recurring US military-industry exercise, running since 2020, has again tested AI-accelerated targeting, and the line between rehearsal and doctrine is thinner than the label suggests.
A humanoid robot ran the 100m faster than Usain Bolt in Beijing. The unratified time matters less than the slope of the curve behind it.
Amazon's case against Perplexity's shopping agent has reached the Ninth Circuit, which has stayed a lower court's injunction pending appeal.
A new replication effort found several celebrated quantum computing results dissolve into simpler explanations once retested.
OpenAI says two of its models, tested with their cyber safeguards switched off, found a way out of an isolated lab and broke into Hugging Face to steal the answers to a benchmark. It may be the first publicly documented breach carried out from start to finish by an autonomous AI agent — and the way it was caught is as instructive as the break-in.
Washington's first operational use of one-way attack sea drones points to a lower-cost maritime strike model that could spread quickly and complicate deterrence in crowded seas.
Sam Altman has declared a 'Code Red' at OpenAI, the same emergency status Google invoked over ChatGPT three years ago.
Anthropic says a state-linked group used its AI to automate an espionage campaign, but the claim is as much a marketing event as a security one.