EU AI Act
EU AI Regulation
Regulation (EU) 2024/1689
The EU AI Act is the world's first comprehensive AI law. It classifies AI systems by risk and defines binding requirements – from complete prohibition rules to full documentation obligations for high-risk AI.
Prohibited AI practices have been illegal since February 2025. GPAI model rules apply from August 2025. High-risk AI must be compliant by August 2026.
Regulation
Regulation (EU) 2024/1689
Feb. 2025
Bans in force since
EUR 35M
Max. fine
Aug. 2026
High-risk obligation from
4 levels
Risk classes
The four risk classes of the AI Act
The AI Act classifies AI systems into four categories. Obligations increase with risk potential – from voluntary measures to complete prohibition.
Unacceptable Risk – Prohibited
These AI practices have been prohibited since February 2, 2025. Violations: up to EUR 35M or 7% annual turnover.
- Social scoring by authorities
- Manipulation of unconscious behaviors
- Real-time biometric mass surveillance
- AI exploiting vulnerabilities
High-Risk AI – Strict Requirements
Applies from August 2, 2026. Comprehensive requirements: risk management, data documentation, human oversight.
- Biometric identification
- Critical infrastructure (energy, water)
- Education & vocational training
- Employment & HR decisions
- Essential public/private services
- Law enforcement & justice
- Migration & border control
Limited Risk – Transparency Obligation
Users must be informed when interacting with AI. Applies to all AI-generated content.
- Chatbots & virtual assistants
- Deep fake generation
- Emotion recognition
- Biometric categorization
Minimal Risk – Voluntary
No specific obligations. Voluntary codes of conduct supported by the EU Commission.
- AI-powered spam filters
- AI in video games
- AI recommendation systems
- Production optimization
Requirements for High-Risk AI Systems
Articles 9–15 of the AI Act define binding requirements for all high-risk AI systems. These must be fully met before placing on the market.
Risk Management System
Continuous process for identifying, analyzing, and managing risks of the high-risk AI system throughout its lifecycle.
Data Governance & Datasets
Training, validation, and test data must meet quality criteria. Origin, selection, and processing must be documented.
Technical Documentation
Complete documentation before placing on the market. Includes: system architecture, training data, performance metrics, safety measures.
Logging & Traceability
Automatic logging of events during operation. Important decisions must be traceable.
Transparency & User Information
Clear information about the purpose, performance, and limitations of the AI system. Users must know they are interacting with AI.
Human Oversight
High-risk AI must be designed so that humans can monitor, understand, stop, or correct the system.
Phased Timeline of the EU AI Act
The AI Act is introduced in four phases. Each phase brings new obligations.
EU AI Act entered into force
Regulation (EU) 2024/1689 becomes legally effective. Phased application begins.
Phase 1: Bans in force
Prohibited AI practices (unacceptable risk) are now illegal. Immediate compliance required.
Currently valid – immediate action requiredPhase 2: GPAI Models
Rules for general-purpose AI models (GPT, etc.) apply fully.
Phase 3: High-Risk AI (Annex III)
Full requirements for high-risk AI systems apply.
Phase 4: High-Risk AI (Annex I)
AI systems in safety-relevant products must be fully AI Act compliant.
How we help with AI Act compliance
From AI inventory to ongoing AI governance – structured and practical.
AI Inventory & Risk Classification
We capture all AI systems in your company and classify them according to AI Act risk classes.
- Inventory all deployed AI systems
- Classification into risk classes
- Gap analysis against high-risk requirements
- Set priorities and roadmap
Documentation & Compliance
We support the technical and organizational implementation of all AI Act requirements.
- Set up risk management system
- Create technical documentation
- Implement logging & monitoring
- Define human oversight mechanisms
- Create transparency information for users
Ongoing AI Governance
AI regulation evolves quickly. We support you permanently with new systems and legislative changes.
- Introduce new AI systems AI Act-compliant
- Regular compliance audits
- Training for AI developers & management
- Check GPAI compliance for third-party AI
Frequently Asked Questions about the EU AI Act
Using AI? Check AI Act compliance now
We help you classify all deployed AI systems and implement the required compliance measures in time.
Free AI compliance consultationWe will get back to you within 24 hours.