Table of Contents
The European AI Regulation (AI Act) came into force on August 1, 2024. With its 113 articles and numerous annexes, it may seem intimidating. Here is a summary to understand the essentials in 10 minutes.
1. What is the AI Act?
The AI Act (Regulation EU 2024/1689) is the world's first comprehensive AI legislation. It applies to any AI system marketed or used in the European Union, regardless of the provider's origin.
Its goal: to ensure that AI is safe, transparent, and respectful of fundamental rights, while promoting innovation. It adopts a risk-based approach: the higher the risk, the stricter the obligations.
Any company that develops, deploys, or uses AI systems in the EU, including non-European companies targeting the European market.
2. The Risk Pyramid
The AI Act classifies AI systems into 4 risk levels. This classification determines your obligations:
The majority of AI systems (chatbots, recommendations, translations) fall into the "minimal" or "limited" categories. Only systems affecting sensitive areas (HR, health, credit) are classified as "high risk."
3. What is Prohibited (Article 5)
Certain uses of AI are completely prohibited in the EU:
Subliminal Manipulation
Techniques that exploit psychological vulnerabilities to manipulate behaviors.
Biometric Surveillance
Real-time facial recognition in public spaces (except for security exceptions).
Social Scoring
Evaluating citizens based on their social behavior like in China.
Predictive Scoring
Crime prediction based solely on profiling.
4. High-Risk Systems (Annex III)
A system is classified as "high risk" if it operates in one of these areas:
- Recruitment and HR management (CV screening, performance evaluation)
- Access to education (grading, academic guidance)
- Credit assessment and financial scoring
- Essential public services (benefits, welfare)
- Law enforcement (identification, risk assessment)
- Medical devices (diagnosis, treatment)
These systems must undergo a conformity assessment before market placement and comply with strict requirements for documentation, transparency, and human oversight.
5. Your Main Obligations
Based on your risk level, here are your obligations:
| Obligation | Article | Applies to |
|---|---|---|
| Risk management | Art. 9 | High risk |
| Data governance | Art. 10 | High risk |
| Technical documentation | Art. 11-12 | High risk |
| Human oversight | Art. 14 | High risk |
| User transparency | Art. 50 | All (if interaction) |
6. Timeline and Penalties
The AI Act enters into application gradually:
August 2024: Entry into force
February 2025: Prohibitions applicable
August 2026: All obligations
- Prohibited practices: up to €35M or 7% of global turnover
- High-risk non-compliance: up to €15M or 3% of turnover
- False information: up to €7.5M or 1% of turnover
Assess your compliance in 5 minutes
Our automated audit determines your risk level and specific obligations.
Start free auditThe AI Act represents a major turning point for the AI industry in Europe. Even if most companies are not affected by the heaviest obligations, it is crucial to verify your situation now to anticipate 2026.
