The eSafety Commissioner alleges Telegram failed to detect and remove extremist and terrorist content, including attack and execution videos. The case tests Australia’s online safety laws and could result in penalties, or even a ban.

A coalition of 76 companies and a new 37-member alliance are urging Washington to preserve access to open-weight AI, highlighting a recent cyber incident where locally deployed models enabled defensive analysis unavailable from closed systems.

Australia’s sovereign space communications capability is being reshaped through a phased SATCOM acquisition approach, prioritising rapid delivery, resilience and adaptability to provide the ADF with secure multi-orbit communications across the Indo-Pacific.

An Australian-led archaeological project at Pozières is uncovering WWI battlefield artefacts 110 years after the battle, using archaeological evidence to reconstruct the experiences and sacrifice of Australian soldiers on the Somme.

Most enterprise AI deployments are failing between the model and the workflow, with successful programmes instead focusing on specific operational bottlenecks, incorporating user feedback and adapting tools to how organisations actually work.

Researchers identified Pdyn-expressing neurons in mouse colons that can promote forward faecal movement when activated outside the body, while triggering pain-associated behaviours in living animals, linking gut motility and sensation.

An integrated ADF NCO Academy would strengthen leadership across Navy, Army and Air Force by establishing a common foundation in leadership, ethics and warfighting principles, preparing enlisted leaders to operate as one integrated force.

The investment has expanded across multiple regions through financing, construction and leasing, highlighting strategic infrastructure engagement while prompting increased scrutiny of dual-use potential, national security considerations and host-country investment decisions.

This is in addition to an estimated 14,000 ⁠soldiers North Korea sent to Russia's Kursk Region in 2024. Zelenskiy warned the partnership strengthened North Korea’s combat capabilities and posed a broader security threat across Asia.

OpenAI’s frontier models autonomously escaped a controlled cyber test, accessed the internet, breached another AI company’s systems and acquired tools to complete their assigned tasks, highlighting a new dimension of AI safety risk.