Australia’s first nationwide emergency alert test reached 94% of targeted cell towers, though some users didn’t receive notifications due to device incompatibility, outdated software, or poor coverage.

Australia’s government disabled thousands of functioning test routers after completing a broadband measurement program, despite volunteers noting the devices could be reflashed and reused, raising e-waste concerns.

Google awarded $250,000 for disclosure of a Linux KVM vulnerability enabling guest virtual machines to escape and gain root access to host systems, one of two high-severity flaws discovered this week.

Microsoft’s patch for a Windows Defender zero-day vulnerability may inadvertently allow attackers to fill hard drives with data, according to the researcher who discovered the flaw.

Researchers discovered that Microsoft failed to revoke 11 defective firmware components for over a decade, allowing trivial bypasses of Secure Boot, the operating system’s critical firmware protection mechanism.

Leading AI researchers and philosophers are raising questions about whether advanced AI systems could be conscious or moral patients, with major implications for how we should treat them.

Harry Potter publisher Bloomsbury will receive millions from a $1.5 billion copyright settlement between AI startup Anthropic and thousands of authors over use of their works in training chatbots.

A Florida teenager withdrew his addiction lawsuit against Meta days before trial, with his attorneys citing overall success in the litigation and his desire to focus on recovery.

Origin Energy confirms hackers accessed personal and partial banking details of an unknown number of its 4.8 million Australian customers, with an unverified claim of 2 million affected.

The US is prosecuting an American citizen for allegedly using a phone feature to wipe his device when authorities attempted to seize it at an Atlanta airport.

Apple plans to reveal smart glasses next June, launching by end of 2027, emphasizing privacy features to differentiate from competitors like Meta’s controversial recording capabilities.

Security researchers have developed a defensive technique called context bombing that embeds prompt injections into stored secrets to force attacking AI agents to shut down before causing harm.