HomeFines & Penalties
Regulatory Risk
EU AI Act Fines & Penalties
Non-compliance with the EU AI Act can result in fines of up to €35 million or 7% of global annual turnover — whichever is higher.
€35M
or 7% turnover
Prohibited AI practices
€15M
or 3% turnover
High-risk & transparency
€7.5M
or 1% turnover
Misleading information
Fines apply to organisations of all sizes. SMEs and start-ups receive the lower of the two amounts (percentage or fixed). Penalties are imposed by national market surveillance authorities, which are active from August 2025. Source: Articles 99–101, Regulation (EU) 2024/1689.
The Three Penalty Tiers
The EU AI Act uses a tiered penalty system — the most serious violations carry the highest fines.
Tier 1 — Prohibited AI Practices
Article 5 + Article 99(3)
CRITICAL
Maximum Fine
€35,000,000
Or % of Turnover
7%
of global annual turnover
or 7% of global annual turnover
(whichever is higher)
Example Violations
- Subliminal manipulation of human behaviour
- Exploitation of vulnerabilities (age, disability, social situation)
- Social scoring by public or private entities
- Real-time remote biometric identification in public spaces (with limited exceptions)
- Emotion recognition in workplaces and educational institutions
- Biometric categorisation inferring race, political opinions, religion
- Predictive policing based solely on profiling
Tier 2 — High-Risk AI & Transparency Obligations
Articles 16, 22–26, 50 + Article 99(4)
HIGH
Maximum Fine
€15,000,000
Or % of Turnover
3%
of global annual turnover
or 3% of global annual turnover
(whichever is higher)
Example Violations
- Failure to establish a risk management system (Art. 9)
- Non-compliant data governance and training data practices (Art. 10)
- Missing or inadequate technical documentation (Art. 11)
- Failure to ensure human oversight capabilities (Art. 14)
- Non-registration in the EU AI Act database (Art. 49)
- Failure to disclose AI-generated content (Art. 50)
- Deployers failing to conduct fundamental rights impact assessments (Art. 27)
Tier 3 — Misleading Information to Authorities
Article 99(5)
MODERATE
Maximum Fine
€7,500,000
Or % of Turnover
1%
of global annual turnover
or 1% of global annual turnover
(whichever is higher)
Example Violations
- Supplying incorrect information to notified bodies
- Providing incomplete responses to national competent authorities
- Misleading market surveillance authorities during investigations
GPAI — General-Purpose AI Model Violations
Article 53–55 + Article 101
HIGH
Maximum Fine
€15,000,000
Or % of Turnover
3%
of global annual turnover
or 3% of global annual turnover
(whichever is higher)
Example Violations
- Failure to provide technical documentation for GPAI models (Art. 53)
- Non-compliance with copyright transparency obligations
- Failure to implement adversarial testing for systemic-risk models (Art. 55)
- Failure to report serious incidents to the Commission (Art. 55)
Aggravating Factors (Article 99(7))
Authorities consider the following when setting the fine amount:
Nature, gravity and duration of the infringement
Number of affected persons and level of damage
Intentional or negligent character of the infringement
Previous infringements by the same operator
Financial benefits gained from the infringement
Degree of cooperation with authorities
Size of the undertaking and market share
Actions taken to mitigate harm to affected persons
When Do Penalties Apply?
ACTIVE
2 Feb 2025
Prohibited AI Practices
Article 5 — violations are enforceable now
ACTIVE
2 Aug 2025
GPAI Models
Articles 53–55 — obligations for GPAI model providers
FROM 2026
2 Aug 2026
High-Risk AI Systems
Articles 6–50 — full obligations for high-risk AI
Act Now
Protect Your Organisation from Fines
Our compliance services help you identify and remediate violations before regulators act.