Legal sources with official primary sources

Printed ·

Skip to main content
Skip to the answer

AI Act Compliance Agent: risk level, requirements, evidence

Describe the system in four choices. The agent classifies the risk level against read articles of Regulation (EU) 2024/1689, shows the requirements and returns an evidence package with a hash. Same choices, same answer, every time.

Share this page

AI Act articles are read in sequence, a clear sign of regulatory use.

Describe the system

AI Act classification

Risk level

High risk

Hash:
a39f2fe47d01f9c8
Articles read:
2026-08-18
Audit-ready run:
32024R1689 · legal-2026-08-25
Articles:
2, 3, 6, 16, 99

Deterministic classification against read articles of Regulation (EU) 2024/1689. Not legal advice.

Claim cards

  1. #claim-1supported

    The system falls within the scope of the AI Act.

    Requirements

    • Assess role and market against Article 2.
    • Use the definitions in Article 3 when describing the system.
    Show evidence

    Hash: 0463476a4f8c8672

  2. #claim-2supported

    The system is classified as high risk and the provider obligations apply.

    Requirements

    • Test the classification against Article 6 and Annex III.
    • Meet the provider obligations in Article 16.
    • Note the penalty levels in Article 99.
    Show evidence

    Hash: 2da32bb66408933c

Evidence chain

  1. 32024R1689 · Articles 2

    Scope

    RegisterOfficial source

  2. 32024R1689 · Articles 3

    Definitions

    RegisterOfficial source

  3. 32024R1689 · Articles 6

    Classification rules for high-risk AI systems

    RegisterOfficial source

  4. 32024R1689 · Articles 16

    Obligations of providers of high-risk AI systems

    RegisterOfficial source

  5. 32024R1689 · Articles 99

    Penalties

    RegisterOfficial source

Agent log

  1. 01 ask , Ask: Four choices are normalised into a question and a run id.
  2. 02 retrieve , Retrieve: Articles are read from the register, no external calls.
  3. 03 claim , Claim: One statement per rule, with requirements and articles.
  4. 04 verify , Verify: Every statement is hashed against the build version.
  5. 05 decide , Decide: Classification, or abstention when the basis is thin.

agent-protocol/1.0.0 · run a39f2fe47d01f9c8

The same answer as a machine

The claim surface returns the same statements with evidence, requirements and hash. The protocol describes the steps an engine runs: ask, retrieve, claim, verify, decide.

Ready-made pack

AI Act compliance pack

€199, one-time purchase

The AI Act as working material: the articles, the high-risk requirements, the documentation duties and a structure for the risk assessment.

Next step

Three ways to put the register to work in your own practice.

Start with your task