Microsoft disrupts AI-assisted platform that compromised 12,000 accounts

A September 2026 Ars Technica security feed highlights active and emerging cyber threats, including AI-assisted account compromise, exploitation of AI-agent privileges, supply-chain attacks, ClickFix social engineering, Chrome and Windows exploit kits, large-scale data breaches, BGP hijacking, malicious consumer devices, and insecure AI-agent behavior. The most immediate risks involve mass account compromise, zero-day exploitation, malware delivery, and attacks against corporate software supply chains. No specific CVE identifiers are provided in the feed metadata.