Is your business in scope of the EU AI Act, and which risk tier applies? A step-by-step walkthrough built directly from Regulation (EU) 2024/1689's own text, drawing on Articles 2, 3, 5, 6, and 113.
The EU AI Act (Regulation (EU) 2024/1689) does not apply uniformly. Applicability depends on an organisation’s role and the nature of its AI use, and even where the Regulation does apply, its obligations are being phased in over several years rather than taking effect at once (Article 113). Establishing an organisation’s position therefore requires two distinct assessments: whether the Regulation applies at all, and, where it does, which of its obligations are currently enforceable. The sequence below addresses both, working through scope, role, prohibited or high-risk status, and the current state of enforceability. Select an answer at each stage to proceed; earlier answers can be revisited at any point.
Misclassification carries a genuine cost in either direction. Under Article 99(3), engaging in a practice prohibited by Article 5(1) while treating it as permissible exposes an organisation to fines of up to €35 million or 7% of global annual turnover, whichever is higher. Under Article 99(4), failing to meet an applicable high-risk obligation under Article 6 carries fines of up to €15 million or 3% of turnover. Conversely, treating a use as high-risk when it is not results in unnecessary expenditure on conformity assessment and technical documentation for an obligation that was never triggered. Both errors typically arise the same way: relying on a secondary summary of the Regulation instead of checking the Article itself, which is why each stage above links directly to the relevant provision on EUR-Lex.