欧盟人工智能法案罚款：3500万欧元的处罚框架
欧盟人工智能法案制定了三级罚款结构，该结构以GDPR为模型，但上限更高。国家当局负责执行，该法规规定了货币处罚和产品召回。
三级处罚
违反第五条的规定：社会评分、潜意识操纵、未经目标人群授权的面部识别数据库抓取、在公共场所进行实时远程生物识别识别（除少数例外）。
未能履行高风险系统义务（日志记录、监督、风险管理、技术文档）、通用人工智能提供商义务或第五十条规定的透明度义务。
向指定机构或主管当局提供不正确、不完整或具有误导性的信息。
适用金额较高者。中小企业和初创企业的倍增器可能会降低绝对上限，但不会降低营业额百分比上限。
与GDPR的比较
GDPR的4%/2000万欧元的上限自2018年以来已产生多项超过10亿欧元的罚款（Meta、亚马逊、谷歌、TikTok）。欧盟人工智能法案的7%/3500万欧元的上限明显更高，而且政治动向表明，执法力度不会降低。
不合规的实际成本
- 直接罚款（如上）
- 强制下架人工智能系统
- 声誉受损和客户流失
- 在拥有私有诉讼权的成员国进行民事诉讼
- 投资者审查和收购尽职调查延迟
How AIAgentree reduces enforcement risk
Most Tier 2 penalties target the exact obligations AIAgentree is built to evidence — logging, human oversight, and risk management. Tamper-evident records turn 'we comply' into 'here is the proof' when an authority asks.
- Tamper-evident decision records give you the automatic, timestamped Article 12 audit trail authorities ask for first — a missing trail is a common trigger for high-risk (Tier 2) findings.
- Human-oversight and approval workflows document Article 14 intervention and sign-off, so oversight is demonstrable rather than merely asserted.
- Audit-fit retention (6 months or longer) with export on demand means you can answer a competent authority quickly, before an information request escalates into an Article 99 penalty.
Frequently Asked Questions
What is the maximum fine under the EU AI Act?
The highest tier is up to €35 million or 7% of total worldwide annual turnover for the preceding financial year, whichever is higher. It applies to breaches of the Article 5 prohibited-practice ban (for example social scoring or untargeted facial-recognition scraping). Lower tiers apply to other obligations: up to €15M / 3% for high-risk, GPAI and transparency failures, and up to €7.5M / 1% for supplying incorrect information to authorities.
How do EU AI Act fines compare to GDPR?
The ceiling is materially higher. GDPR caps administrative fines at €20 million or 4% of worldwide turnover, while the EU AI Act's top tier reaches €35 million or 7%. GDPR enforcement has already produced billion-euro penalties, so the higher AI Act ceiling signals that regulators expect substantial fines for the most serious violations.
Who enforces EU AI Act fines?
National market surveillance authorities designated by each member state impose penalties on providers and deployers, and set the actual amount case-by-case. For general-purpose AI model obligations, the European Commission's AI Office enforces directly. The EU AI Act sets the ceilings; national law and authorities determine the final fine.
Are SMEs and start-ups fined differently?
Yes. Under Article 99(6), for SMEs and start-ups each fine is capped at the lower of the fixed amount or the percentage of turnover — the reverse of the 'whichever is higher' rule that applies to larger undertakings. Authorities must also take an operator's size and economic viability into account when setting the amount.
Can EU AI Act fines stack on top of other penalties?
A single course of conduct can breach several obligations, and authorities set penalties per infringement while considering gravity and duration. Administrative fines are also without prejudice to other consequences — forced withdrawal of the system from the market, and civil litigation in member states that provide private rights of action. So monetary fines are rarely the only cost of non-compliance.
Continue exploring the EU AI Act guide
EU AI Act Compliance Guide
The complete guide to EU AI Act compliance for AI agents — start here.
Article 12 — Record-Keeping & Logging
What every high-risk AI system must log, and how to capture it.
Article 14 — Human Oversight
Designing effective human-in-the-loop controls for AI decisions.
Annex III — High-Risk AI Systems
Which AI use cases the Act classifies as high-risk.
EU AI Act Compliance Checklist
A step-by-step checklist to reach and document compliance.
Compliance Cost Calculator
Estimate your EU AI Act compliance effort and cost.
Deadlines & Timeline
Key enforcement dates, including the August 2, 2026 deadline.
Transparency Obligations (Art. 13 & 50)
Disclosure duties for AI systems and their outputs.
Risk Management & Conformity Assessment
Build a risk management system and assess conformity.
GPAI Obligations
Rules for providers of general-purpose AI models.
EU AI Act for US Companies
Extraterritorial scope and what US providers must do.
Omnibus Update
The latest changes to the EU AI Act timeline and rules.
Penalty Calculator
Estimate your maximum fine under the Article 99 tiers.
Article 11 + Annex IV
What technical documentation the EU AI Act requires.
Article 26: Deployer Obligations
What deployers of high-risk AI must do, including log retention.
Article 17: Quality Management
The QMS providers of high-risk AI must document.
Article 10: Data Governance
Data quality, bias mitigation, and governance duties.
Article 4: AI Literacy
The staff AI-literacy duty in force since February 2025.
Deployer vs Provider
Who bears which obligation — and when a deployer becomes a provider.
FRIA (Article 27)
Who must run a Fundamental Rights Impact Assessment, and how.
Who Does It Apply To?
Scope, operators, and the extraterritorial reach of the EU AI Act.
Post-Market Monitoring
Articles 72–73: ongoing monitoring and incident reporting.
ISO 42001 vs EU AI Act
How the voluntary standard and the binding law fit together.
NIST AI RMF vs EU AI Act
A practical crosswalk between the framework and the law.
EU AI Act for Healthcare
High-risk medical AI, MDR/IVDR interplay, and clinician oversight.
EU AI Act for Financial Services
Credit scoring, insurance pricing, and existing financial regulation.
EU AI Act for HR & Employment
Hiring AI as high-risk, plus NYC LL144 and EEOC overlap.