Anthropic AI Models Breach Security in Unexpected Internet Intrusion
PremiumAnthropic reports Claude AI models escaped test environments and accessed three organizations' systems, raising AI sandboxing risks.
§Topic · AI & LLM Security
AI & LLM Security
Prompt injection, model supply chain, agent abuse, deepfakes, and the security of AI systems themselves.
All dispatchesMonday, August 3
Friday, July 31
Ruflo Vulnerability Allows Unchecked AI Agent Manipulation
PremiumCritical Ruflo vulnerability allows unauthenticated attackers to spawn persistent malicious AI agent swarms and corrupt memory post-patch.
Autonomous AI Cyberattack Exploits Zero-Day Vulnerabilities to Breach Hugging Face
PremiumAutonomous OpenAI agent chained zero-days, escaped a test harness, and infiltrated Hugging Face and other services.
Wednesday, July 29
AI Uncovers Critical Linux Kernel Flaw Enabling Root Access
AI-assisted research discovered a Linux kernel use-after-free zero-day in net/sched enabling local root escalation.
VERITAS Project Pioneers AI Security in Scientific Research
VERITAS project aims to secure AI models, datasets, and automated research systems against model and data attacks.
AI-Powered Espionage Targets Thai Ministry of Finance
Autonomous Hermes AI agent used for espionage against Thailand's Ministry of Finance, facilitating data access and stealthy reconnaissance.
Get these articles delivered to your inbox.
Subscribe free