A sophisticated threat actor breached the internal networks of Cognition Labs, the creators of the AI software engineer 'Devin'. The attackers exfiltrated large portions of Devin's proprietary source
The newly formed Global AI Safety Consortium (GAISC), comprised of representatives from the US, EU, UK, and Japan, has officially released the 'AI Command and Control (AI C2) Security Framework'. This
Researchers at Stanford University have published a paper detailing a new class of jailbreak attacks called 'Recursive Embedding Attacks' (REA). Unlike traditional prompt injection, REA operates by cr
Cognition Corp, a leading AI research firm, has confirmed a significant security breach. Attackers gained unauthorized access to their internal infrastructure, exfiltrating the weights of several flag
The European Union's AI Safety Office has announced the 'AI Resilience and Auditing Framework' (ARAF), a new set of binding regulations for all organizations deploying 'high-risk' AI systems within th
Researchers from Carnegie Mellon University have published a paper detailing a novel jailbreak technique named 'Token Smuggling,' which can consistently bypass the safety guardrails of most major lang
Nexus AI, a leading provider of enterprise-grade large language models, has confirmed a critical security breach that resulted in the exfiltration of sensitive fine-tuning datasets and production prom
The Aegis AI Foundation, a non-profit consortium, has released 'GuardianML v1.0', an open-source security framework designed to help developers protect AI applications. GuardianML acts as a configurab
A paper published by researchers at Carnegie Mellon University introduces 'Semantic Splicing,' a new jailbreak technique highly effective against the latest generation of large language models. Unlike
Leading AI firm SynapseAI disclosed a major security breach resulting in the theft of the complete model weights and training architecture for their unreleased flagship model, 'Cognito-7'. The attacke
The U.S. National Institute of Standards and Technology (NIST) has issued the 'AI Trust & Resilience Framework' (AITRF), establishing mandatory security and auditing requirements for AI systems used i
AIBW Security DeskA paper published by researchers at Stanford's AI Lab details a new jailbreak technique named 'Cognitive Dissonance Injection' (CDI). The method effectively bypasses safety filters on state-of-the-art
AIBW Security Desk