Question 1

What is an AI risk score?

Accepted Answer

An AI risk score is a quantified 0–100 safety rating that evaluates an AI model across four pillars: regulatory compliance, security & vulnerability, ethical bias, and data privacy. Higher scores indicate lower risk. Global AI Risk provides these scores for 500+ leading AI models.

Question 2

How does Global AI Risk assess AI models?

Accepted Answer

Global AI Risk uses a proprietary four-pillar methodology aligned with EU AI Act, NIST AI RMF, and ISO 42001 standards. Each model is evaluated on regulatory compliance, security vulnerabilities, ethical bias, and data privacy using automated scanning and expert analysis.

Question 3

Which AI models does Global AI Risk cover?

Accepted Answer

Global AI Risk currently covers 500+ AI models including GPT-4o, Claude 3.5, Gemini 2.0, Llama 3.1, Mistral, and many more. New models are continuously discovered and added through our automated scanning system.

Question 4

What compliance frameworks does Global AI Risk support?

Accepted Answer

Global AI Risk maps risk assessments to six major frameworks: EU AI Act, NIST AI RMF, ISO 42001, OECD AI Principles, Singapore FEAT, and Canada AIDA. Compliance certificates are available for enterprise subscribers.

Trust centre