generative ai security

Harmonic Security’s analysis of 22.4 million generative‑AI prompts in 2025 reveals a stark concentration of enterprise data exposure: six applications account for more than 92% of measured potential leakage, and ChatGPT alone drives roughly 71% of those exposures despite representing less than...

A single click on a seemingly harmless Copilot link, a steady stream of employees pasting sensitive text into public chatbots, and consumer AI apps moving from conversation to commerce — together these developments expose a brittle set of trust boundaries in today’s generative-AI ecosystems and...

Microsoft’s new e-book on generative AI security lands at a pivotal moment: defenders are racing to embed AI into detection, response, and automation pipelines even as adversaries weaponize the same technology to scale phishing, deepfakes, and adaptive malware. The guide — 5 Generative AI...

A chilling new wave of cyber threats has emerged at the intersection of artificial intelligence and enterprise productivity suites, exposing deep-rooted vulnerabilities in widely adopted platforms such as Microsoft 365 Copilot. Among the most unsettling of these discoveries is a “zero-click” AI...