File type: PDF (34 pages, ~217 KB) · Status: demo delivery, fictional data, not binding · Last generated: 2026-08-31
Get the full report
Inline preview below. Download for the complete version — all 9 points, source-cited, Ed25519-signed.
Inline preview
First page rendered inline. Click to open it in the full viewer.
The 9 points of Annex IV
Every point is filled in in the sample report with fictional but realistic data.
- Point 1. A general description of the AI system — intended purpose, provider, version, intended use and intended users.
- Point 2. The elements of the system and the development process — architecture, design choices, training data, validation and testing.
- Point 3. Monitoring, functioning and control — capabilities, limitations, expected accuracy and logging.
- Point 4. The appropriateness of the performance metrics — why exactly these measures are the right ones for this system.
- Point 5. The risk management system pursuant to Article 9 — identified risks, controls and residual risk.
- Point 6. Relevant changes made through the lifecycle — version history and what each change meant.
- Point 7. Harmonised standards applied — or the solution chosen where no standard has been applied.
- Point 8. The EU declaration of conformity pursuant to Article 47.
- Point 9. The post-market monitoring system pursuant to Article 72.
Verification
The sample delivery carries a signed Evidence Manifest. Verify the Ed25519 signature at /verify/e1f36f34…018e — or offline with Python against the public key at /.well-known/powerquant.pub. Detailed description at How the evidence works.
Generated by the PowerQuant Module 1 demo — source-cited, Ed25519-signed. Verify the signature at /verify/{hash} or offline against /.well-known/powerquant.pub.