Major AI players unite to create a

A consortium of major AI companies called the Frontier Model Forum, backed by Anthropic, Google, Microsoft, and OpenAI, has released a draft framework for rating the severity of AI jailbreak attacks. The 'Jailbreak Rating Framework' is designed to be a standard scoring system, similar to the CVSS in cybersecurity, to classify how an AI model's safety guardrails can be bypassed. It assesses techniques based on the dangerous new capabilities they unlock and how easy they are for attackers to deploy. The goal is to give developers, security teams, and regulators a common, evolving metric to prioritize fixes and assess risks as models improve. This move signals the industry's growing recognition that AI safety vulnerabilities need systematic, transparent gradingâjust like traditional software bugs.