The EU AI Act has teeth — fines are higher than under GDPR and can hit companies hard. In this lesson, you'll learn what penalties apply, who enforces them, and how to file complaints.
Fine Levels
The AI Act defines three fine levels — depending on the severity of the violation:
Level 1: Prohibited AI Practices
Up to €35 million or 7% of global annual turnover (whichever is higher)
Applies to: Use of prohibited AI systems (social scoring, manipulative AI, etc.)
For comparison: GDPR maximum €20 million or 4% of turnover
Level 2: High-Risk Violations
Up to €15 million or 3% of global annual turnover
Applies to: Non-compliance with High-Risk system obligations
Examples: Missing documentation, no risk management, missing logging
Level 3: False Information
Up to €7.5 million or 1% of global annual turnover
Applies to: False or misleading information to authorities
Examples: False conformity declaration, concealment of risks
Special Rules for SMEs and Start-ups
For small and medium enterprises (under 250 employees) and start-ups, reduced fines apply — the lower of the two values (absolute amount or turnover percentage) is applied.
Competent Authorities
EU Level
EU AI Office: Central coordination body, responsible for GPAI models
AI Board: Advisory body composed of member state representatives
National Level
Each member state designates:
Market surveillance authority: For monitoring the AI Act
Notifying authority: For designating conformity assessment bodies
Germany: The Federal Network Agency (BNetzA) is designated as the national market surveillance authority. Details are regulated in the national implementing law.
Complaint Mechanism
Affected persons can file complaints with:
The competent national market surveillance authority
Consumer organizations and associations
The procedure resembles the GDPR complaint mechanism:
File a complaint (informal submission possible)
Authority examines the matter
Authority can order measures (adaptation, recall, fine)
Legal recourse before courts possible
Liability and Damages
Beyond fines, civil law consequences also apply:
AI Liability Directive (in preparation): Eased burden of proof for affected parties
Product Liability Directive: AI systems are considered products — manufacturers are liable
Contractual liability: Toward customers receiving defective AI products
How to Avoid Penalties
Measure
Risk Reduction
AI inventory with risk classification
Foundation for everything
Create and maintain documentation
Proof of compliance
Implement monitoring
Early warning system for violations
Train employees
Prevention of unintentional violations
Involve legal counsel
Expertise for borderline cases
Remember: The fines are deterrent — but the real danger is reputational damage and market bans. An AI system that must be withdrawn from the market costs more than any fine.
📝
Quiz
Question 1 of 3
Wie hoch ist das maximale Bußgeld für den Einsatz verbotener AI-Systeme?