Self-Assessment
Classify an AI system two ways: a quick guided decision tree that follows the law's order, or a full 10-point audit of every risk factor. Unsure about the levels? Read the AI-risk framework.
Assessment & Classification
Is your AI application compliant?
Answer a few questions about your AI system to get a rough risk classification. This is a simplified guide; always consult a legal or compliance expert for formal classification.
Step 1 of 3 · Art. 5
Does the system use a prohibited practice?
Manipulation that causes harm, exploiting vulnerabilities, social scoring, predicting crime from profiling, emotion recognition at work, or untargeted scraping of faces.
Finance examples
- Upselling high-fee products to customers flagged as financially vulnerable
- Scoring customers on unrelated social behaviour to deny accounts