OWASP AIVSS targets agentic AI risk

The report explores OWASP’s new AI Vulnerability Scoring System (AIVSS) as an extension to CVSS to address AI-specific risks that traditional frameworks miss, such as model/data-centric issues, agentic behaviors, and governance needs. Experts argue AIVSS can help prioritize AI vulnerability response and align with regulatory expectations, while cautioning about potential shortcomings like arbitrary weighting and the need for clear remediation guidance. The roadmap includes industry-specific calculators, education and certification, mobile AI considerations, and broader coverage beyond agentic AI.