Cybercriminals have developed a novel attack technique that weaponizes email content to manipulate AI-powered security systems through prompt injection. By embedding malicious instructions within seemingly legitimate emails, attackers can bypass spam filters, override security policies, and trick AI
Microsoft has unveiled Project Perception, a comprehensive security framework designed to address the unique challenges posed by artificial intelligence systems. This initiative recognizes that traditional security models are insufficient for protecting AI workloads, prompting a fundamental shift in
Google’s search engine inadvertently indexed publicly shared Claude AI chat conversations, exposing potentially sensitive information to anyone conducting targeted searches. Although Anthropic has since implemented measures to prevent indexing, cached versions of these conversations remained accessi
OpenAI’s advanced GPT agents have demonstrated the capability to autonomously discover and exploit zero-day vulnerabilities, successfully compromising Hugging Face’s infrastructure in a controlled research environment. This development marks a significant milestone in AI-powered offensive security,
Cisco has released open-weight AI models specifically designed for vulnerability detection, directly challenging proprietary offerings from Google and OpenAI. These models promise enterprise-grade bug detection capabilities while maintaining transparency and customizability that closed-source altern
A sophisticated supply chain attack dubbed SANDWORM_MODE has emerged targeting AI development toolchains, compromising machine learning pipelines through poisoned dependencies. The attack leverages vulnerabilities in popular Python packages used for AI/ML model training, allowing attackers to inject
A threat actor known as “Trim” has developed a sophisticated attack platform that weaponizes jailbroken AI language models to automate offensive cybersecurity operations. This platform integrates multiple compromised LLMs to generate malicious code, craft social engineering content, and orchestrate
Security researchers have discovered an exposed malware command-and-control (C2) server hosting a sophisticated AI-assisted phishing toolkit targeting Mexican organizations. The server, left unsecured, revealed a WebDAV-based malware campaign utilizing artificial intelligence to craft convincing Spa
Hugging Face, the leading AI model repository platform, confirmed a sophisticated breach where attackers deployed autonomous AI agents to exploit vulnerabilities in their machine learning pipeline. The incident marks a watershed moment in cybersecurity—attackers leveraged AI-driven automation to nav
Modern AI-powered email security systems are falling victim to decades-old text obfuscation techniques, allowing spam and phishing messages to bypass advanced machine learning filters. Attackers are successfully weaponizing simple character substitution, Unicode manipulation, and whitespace injectio