Skip to main content

Artificial Intelligence System Identification Tool Please provide the system details for accurate identification.

Provide a name and brief description of the system you are evaluating. These data remain on your device.

Insert the internal or commercial name of the system.
0/800

Come utilizzare EU AI Act Self-Assessment 2026

Identify the AI System

Enter system name and description to evaluate (chatbot, scoring model, recommendation system, etc.).

Define purpose and usage context

Indicate purpose, sector, and geographic region of deployment: these are the parameters that determine which obligations under the AI Act apply.

Risk Level Classification

Respond to questions on autonomous decision-making, impact on fundamental rights, sensitive data handled and human oversight to determine the Annex III category.

Analyze compliance gaps

The tool highlights missing obligations relative to the assigned risk category and proposes corrective actions prioritized.

Download Final Report

Get a summary report with risk classification, compliance gap and operational recommendations, exportable by email.

Suggerimenti

  • Even limited risk systems have transparency obligations: don't underestimate the "minimum risk" category without verifying it in the wizard.
  • Use third-party GPAI models (OpenAI, Anthropic, Google) always check the technical documentation provided by the provider: it's an obligation of transparency cascade.
  • Repeat the evaluation every time you change the system's purpose: risk classification depends on the application context, not just technology.

Domande frequenti

Is my chatbot covered by the AI Act?

Yes. Chatbots interacting with physical people fall under the transparency obligations of the AI Act (Art. 50). They must inform users they are interacting with an AI system. Deadline is August 2, 2026.

What are Annex III systems?

High-Risk System Categories Listed Under AI Act Regulation Include: Biometrics, Critical Infrastructure, Education, Employment, Essential Services, Countermeasures, Migration, and Justice. Systems in these categories have burdensome obligations (Articles 9-14).

What are the penalties for violating the AI Act?

Up to 7% of global revenue (35M EUR) for prohibited practices, up to 3% for high-risk system violations, and up to 1.5% for GPAI and informational obligations.

What is a GPAI model?

A General-Purpose AI Model (GPAI) is a model with general capabilities that can be adapted to various tasks (e.g., GPT-4, Claude, Gemini). Subject to the obligations of Articles 53-55 as of August 2, 2025.

Does the result of your self-assessment have legal value?

No: It is an educational and guidance tool, not a certified compliance assessment. For high-risk systems, formal evaluation with a specialized AI Act lawyer is still required.