Administrative Fine

Official/legal basis. The EU Artificial Intelligence Act requires Member States to provide for administrative fines as an enforcement measure for infringements of the Regulation and sets maximum categories and amounts for those fines (for example, up to EUR 35,000,000 or up to 7% of worldwide annual turnover for the gravest infringements). The Act also prescribes that the amount of any administrative fine be determined taking into account a range of circumstances specific to each case. See: EU AI Act, Article 99 (Penalties) and Article 100 (administrative fines on Union institutions). ([eur-lex.europa.eu](https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng?utm_source=openai))

Context and scope. In the context of the AI Act, an administrative fine is a non-criminal, monetary sanction imposed by a competent national authority (or imposed by national courts where required by national legal systems) for non‑compliance with the Regulation’s obligations, prohibitions or transparency requirements. The Regulation distinguishes categories of infringements (e.g., prohibited AI practices; failures to meet system requirements; providing incorrect or misleading information to authorities) and links them to upper limits for fines. Member States must adopt implementing rules so that enforcement is effective, proportionate and dissuasive across their jurisdictions. ([eur-lex.europa.eu](https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng?utm_source=openai))

Practical implications for businesses and organisations. For providers, deployers and other operators in the AI value chain, administrative fines are a primary financial risk from regulatory non‑compliance: they create direct monetary exposure and reputational risk, may be combined with other corrective measures (e.g., orders to suspend or modify systems), and can scale with an undertaking’s worldwide turnover where provided for in the Act. Businesses therefore must: (i) identify whether their AI systems fall within the Regulation’s scope; (ii) implement the required technical, organisational and documentation measures (risk‑management, record‑keeping, transparency, human oversight, post‑market monitoring); and (iii) cooperate with competent authorities during inspections and investigations to mitigate potential penalties. The AI Act also requires Member States to consider the economic viability of SMEs and start‑ups when setting national penalty rules. ([eur-lex.europa.eu](https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng?utm_source=openai))

Key criteria and procedural considerations. When imposing and sizing administrative fines the AI Act requires decision‑makers to consider all relevant circumstances of the individual case. Typical factors (explicitly listed in the Regulation and aligned with established EU practice for administrative fines such as under the GDPR) include:

• the nature, gravity and duration of the infringement and its consequences (including the number of affected persons and level of damage);
• the degree of responsibility of the operator and the technical and organisational measures it implemented;
• whether the operator cooperated with authorities or took steps to mitigate harm;
• previous infringements or fines concerning the same operator;
• the size, turnover and market share of the offending operator (relevant where percentage‑based caps apply); and
• other aggravating or mitigating circumstances (e.g., intentionality or negligence).

These factors reflect the AI Act’s prescribed approach to ensure fines are effective, proportionate and dissuasive, and are consistent with the EU’s prior administrative‑fine framework (notably the criteria found in Article 83 GDPR). Competent authorities must also respect procedural safeguards and judicial remedies under Union and national law. ([eur-lex.europa.eu](https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng?utm_source=openai))

Examples and cross‑references. Examples of situations that may attract administrative fines under the AI Act include: deploying an AI system that falls within a prohibited practice (subject to the highest tier of fines), operating a high‑risk AI system that fails to meet mandatory requirements (intermediate tier), or supplying false or misleading information to a notified body or national competent authority (lower tier). For related concepts see: penalties, market surveillance, post‑market monitoring, notified bodies and judicial remedies under the Regulation. ([eur-lex.europa.eu](https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng?utm_source=openai))