Scenario: A European AI technology company was found to be non-compliant with certain provisions of the EU AI Act. The regulator is considering…

The EU AI Act establishes a tiered structure for administrative fines. The middle tier, which carries penalties of up to €15 million or 3% of the company's total worldwide annual turnover for the preceding financial year (whichever is higher), applies to non-compliance with specific obligations. This includes the failure of providers to meet the requirements for high-risk AI systems, such as those concerning risk management, data governance, technical documentation, transparency, and human oversight. This penalty level is designed to address significant breaches of the Act's core operational requirements for systems that pose a substantial risk to health, safety, or fundamental rights.

A. Breaches of AI Act prohibitions (e.g., social scoring, manipulative AI) are subject to the highest fines, up to €35 million or 7% of annual worldwide turnover.

C. The supply of misleading information to notified bodies or national authorities is subject to the lowest tier of fines, up to €7.5 million or 1% of annual worldwide turnover.

D. This option is incorrect for two reasons: the question concerns a company, not a Union institution, and a breach of prohibitions would fall under the highest penalty tier (€35M/7%).

1. European Parliament. (2024). Artificial Intelligence Act. P9TA(2024)0138.

Article 71

Paragraph 4: "Non-compliance of the AI system with any of the requirements or obligations under this Regulation

other than those laid down in Articles 5 and 10

shall be subject to administrative fines of up to EUR 15 000 000 or

if the offender is an undertaking

up to 3 % of its total worldwide annual turnover for the preceding financial year

whichever is higher." This paragraph directly covers the obligations for high-risk AI systems.

Article 71

Paragraph 3: Specifies the highest fines (€35 million or 7%) for non-compliance with the prohibited practices listed in Article 5. This refutes option A.

Article 71

Paragraph 5: Specifies the lowest fines (€7.5 million or 1%) for the supply of incorrect

incomplete

or misleading information. This refutes option C.

2. Veale

M.

& Borgesius

F. Z. (2021). Demystifying the Draft EU Artificial Intelligence Act. Computer Law Review International

22(4)

97-112.

Section 4.3 'Enforcement and penalties': This section analyzes the proposed penalty structure

noting that "infringements of most of the obligations relating to high-risk AI systems... are proposed to be subject to fines of up to [the middle tier]." While based on the draft

the final Act retained this tiered structure

confirming the principle. (Note: Specific figures were updated in the final text cited above). DOI: https://doi.org/10.9785/cri-2021-220402