EU
← Back to Examples

Email Spam Filter

Minimal Risk

A machine learning system that classifies incoming emails as spam or legitimate. Operates entirely server-side with no direct user interaction.

Classification Walkthrough

  1. Step 1 — Prohibited practices (Article 5): A spam filter does not perform any prohibited practice. ✗ Not prohibited.
  2. Step 2 — GPAI: A spam filter is a narrow, task-specific system, not a general-purpose AI model. ✗ Not GPAI.
  3. Step 3 — High-risk Pathway A: Not a safety component of any Annex I regulated product. ✗ Not Annex I.
  4. Step 4 — High-risk Pathway B: Email filtering does not fall under any Annex III high-risk category. ✗ Not Annex III.
  5. Step 5 — Limited risk: The spam filter operates server-side and does not directly interact with natural persons, generate synthetic content, perform emotion recognition, or perform biometric categorization. ✗ No transparency triggers.
  6. Result: MINIMAL RISK — No specific obligations under the EU AI Act. The provider may voluntarily adopt codes of conduct (Article 69). Enforcement date: August 2, 2026.
Classify like this