Archived page. This page comes from the previous ALT-F1 website and is kept for reference; its content is not maintained. Original address www.alt-f1.be/ai-act-explained.html, snapshot of in the Wayback Machine. Current services: ALT-F1 home page.

AI Act explained

The original page was an interactive explorer whose tabs, risk tiers and charts were drawn by scripts; only its text is kept here. The AI Act and the other frameworks are explained on the current site at /en/frameworks/.

EU AI Act Explorer

An interactive guide to the world’s first comprehensive AI law.

The EU’s AI Act introduces a harmonized set of rules for the development, deployment, and use of AI systems. Its primary aim is to ensure AI is trustworthy and human-centric by establishing a regulatory approach based on the level of risk an AI system poses to individuals and society. This section outlines the core objectives driving this legislation.

The Four-Tiered Risk Approach

The centerpiece of the AI Act is its classification of AI systems into four risk categories. This structure ensures that regulatory intervention is proportional to the potential for harm. Select a risk tier below to explore its definition, examples, and the corresponding legal obligations.

GPAI & Governance

The Act establishes specific rules for General-Purpose AI (GPAI) models, which have wide-ranging capabilities, and creates a new governance structure to ensure consistent enforcement across the EU. This section details the obligations for GPAI providers and the key bodies responsible for oversight.

Implementation Timeline

The AI Act will not be enforced all at once. It has a staggered implementation schedule to give businesses and member states time to adapt. The chart below visualizes the key milestones, from the initial entry into force to when different parts of the regulation become fully applicable. Hover over the points to see what happens on each date.

Penalties for Non-Compliance

To ensure the rules are followed, the AI Act includes significant financial penalties. Fines are structured based on the severity of the violation and a company’s size, with the highest penalties reserved for the use of prohibited AI practices. The chart below illustrates the maximum potential fines for different types of infractions.