Arrest Warrant Rejected: AI Match Only a Vague Clue for Judge
A district court should restrict the use of facial recognition technology and strengthen the rights of defendants against opaque IT-based investigative tools, as automated search results in criminal prosecution alone do not satisfy legal requirements without robust safeguards and technical transparency.
“Fragnesia”: Next Privilege Escalation Vulnerability in Linux Kernel
Microsoft warns of “Fragnesia,” a new Linux kernel vulnerability (CVE-2026-46300, CVSS 9.9) that represents a variant of the Dirty Frag flaw; it affects the XFRM-ESP subsystem in IPsec and allows attackers to gain root privileges, discovered by William Bowling using an AI-powered tool, with exploit code and patches alr
Apple Investigates macOS Issue Discovered by Claude Mythos
As part of the European Commission-funded Glasswing project, Anthropic is collaborating with Apple to identify security vulnerabilities in macOS, with the Claude Mythos AI model deployed by security firm Calif to exploit two bugs and memory errors to gain elevated access rights; details have not yet been published.
F5 BIG-IP: Quarterly Security Update Closes Multiple Vulnerabilities
F5 has published its quarterly security update closing vulnerabilities in BIG-IP, BIG-IQ Centralized Management, NGINX Plus and NGINX Open Source, with a particularly critical flaw in NGINX allowing unauthenticated code execution via crafted HTTP requests that could enable DoS, privilege escalation and access to protec
Ivanti EPM: Security Vulnerabilities Enable SQL Injection and Privilege Escalation
Ivanti warns of three security vulnerabilities in Endpoint Manager (EPM): an SQL injection flaw in the web console allows authenticated attackers to inject malicious code, improper privilege assignment enables local privilege escalation, and a core server error threatens credentials; all three vulnerabilities reported
