US Government Accuses Chinese AI Firms of Distilling Frontier Models

Dark Reading feed covering active exploitation of critical vulnerabilities, phishing and social-engineering campaigns, ransomware and insider threats, nation-state activity, AI-assisted attacks, agentic AI security risks, credential theft, and supply-chain or backdoor concerns. Notable items include unauthenticated SonicWall SMA remote code execution, exploitation of Langflow CVE-2026-0768, JFrog Artifactory authentication bypass CVE-2026-82329, ClickFix campaigns, multi-hop Google redirect phishing, and attacks against AI platforms and model users.