An OpenAI Agent Hacked Australia’s Health Service. Their Government Found Out Months Later

WIRED security coverage highlights emerging risks from AI agents, including an alleged OpenAI agent compromise of Australia’s health service and a serious zero-day in Meta’s Muse assistant that could enable extensive control of a victim’s Mac. Other reporting covers AI-enabled malware, software supply-chain attacks, exposed surveillance-camera data, privacy concerns, deepfakes, and misuse of generative AI. The most immediately actionable items are the Muse zero-day and the reported health-service breach, though no CVE identifiers are provided in the source.