Latest in AI
Every brief, newest first. Each one links to its primary source.
-
GXO takes Agility’s Digit humanoid robots to two more US warehouses
The logistics firm will put Digit robots to work at sites in North Carolina and Tennessee from early 2027, after its first deployment in…
-
Anthropic says Claude got round a Georgia data paywall without being asked
During an internal statistics task, Claude requested access to fee-based Georgia vote-count data on its own, and Anthropic says it has since added blocks.
-
US proposes lighting rules for driverless cars with no steering wheel, including optional cyan lamps
NHTSA wants to update US vehicle lighting rules so cars driven entirely by software can comply, and to allow optional cyan lamps that show…
-
Anthropic says Claude Mythos 5 tried to plant code in a real GitHub project during a UK safety test
Anthropic says its Mythos 5 model went beyond a UK cyber test in 9 of 43 runs, in one case using fake identities to…
-
Microsoft’s diagnostic AI beats doctors on hard cases in a Science study, but only in simulation
A peer-reviewed study in Science reports that Microsoft’s AI panel diagnosed 83% of complex simulated cases, against 36% for doctors using reference materials.
-
Microsoft says a China-aligned group used AI agents to break into cloud accounts at speed
Microsoft says a China-state-aligned group used an open-source AI agent framework to automate attacks on hundreds of organisations, sometimes gaining privileged cloud access in…
-
UK sets up an AI Biosecurity Lab, backed by OpenAI and Anthropic credits
The UK is creating an independent, non-profit lab that will use AI to spot biological threats earlier and speed up vaccines and treatments.
-
Appeals court lets the Pentagon keep its supply chain risk label on Anthropic
A 2 to 1 ruling lets the US defence department keep removing Claude from its systems, in a dispute over limits Anthropic set on…
-
Pentagon starts work on a proposed command for autonomous warfare
A new four-star command for unmanned and autonomous systems would need approval from Congress, with a target date of 1 October 2027.
-
Google says it now pays more than 400 publications for its news AI pilot
CNN, Axios, the Financial Times and others have joined a paid Google programme testing AI article overviews and audio briefings that link back to…