A paper published by researchers at Carnegie Mellon University has detailed a novel jailbreak technique called 'Semantic Smuggling'. The attack leverages the expanding context windows of modern LLMs t
AIBW Security DeskNexus AI, a leading competitor to OpenAI and Anthropic, has confirmed a catastrophic data breach resulting from a misconfigured public cloud storage bucket. The breach exposed over 30 terabytes of sen
AIBW Security DeskThe U.S. Department of Commerce, in collaboration with NIST, has officially enacted the 'AI Model Accountability Framework' (AMAF). This new regulation moves beyond voluntary industry commitments by l
AIBW Security DeskVocalize AI, a prominent provider of voice synthesis and cloning services for corporate and entertainment clients, disclosed a critical security breach. Attackers exploited an unsecured API endpoint i
AIBW Security DeskA paper published by researchers at Carnegie Mellon University details a new LLM jailbreak technique named 'Contextual Code Injection.' The attack involves embedding obfuscated, malicious commands wit
AIBW Security DeskThe European Union's AI Act has officially entered its enforcement phase, with the EU AI Board issuing its first significant penalties. A major social media company was fined €25 million for failing t
AIBW Security DeskMedSynth, a prominent healthcare AI company specializing in diagnostic models, has confirmed a major data breach exposing the protected health information (PHI) of approximately 2 million patients. Th
AIBW Security DeskResearchers at Carnegie Mellon University have published a paper detailing a novel jailbreak technique named 'LogicLock.' This method embeds complex, multi-step logical reasoning puzzles into prompts
AIBW Security DeskThe non-profit SentinelAI Foundation has launched 'Glaive,' a new open-source framework designed to automate the security testing and red-teaming of Large Language Models. Glaive provides a comprehens
AIBW Security DeskA research paper published by Carnegie Mellon University's CyLab introduces a novel jailbreak technique named 'Chameleon.' The attack embeds hidden, malicious instructions within Unicode glyphs that a
AIBW Security DeskNexusAI, a leading AI development firm, confirmed it suffered a significant data breach originating from a supply-chain attack on a third-party data annotation partner. The breach exposed a repository
AIBW Security DeskThe United States Senate has passed the 'AI Verification, Evaluation, and Reporting for Foundational Integrity' (VERIFY) Act, a landmark piece of legislation aimed at improving AI safety and accountab
AIBW Security Desk