Model Safety

Prisma AIRS 2.0: Comprehensive Security Framework for Agentic AI Systems

Prisma AIRS 2.0 represents a fundamental transformation in how enterprises approach agentic AI security, treating it not as an add-on feature but as a distinct class of identity, data flow, and...

AI Jailbreaks Reveal Critical Security Vulnerabilities in Leading Language Models

Introduction Recent research has unveiled that bypassing the safety mechanisms of advanced AI language models remains alarmingly straightforward. This revelation underscores significant security gaps...

Policy Puppetry: Unveiling a Universal Vulnerability in Large Language Models

Introduction Recent research has unveiled a significant vulnerability in Large Language Models (LLMs), termed "Policy Puppetry." This technique allows adversaries to bypass safety mechanisms across...