One of the main problems of the EU AI Act is its risk-based approach to AI regulation. Under the Act, AI systems are classified by their risk level, and high-risk systems will be subject to the most strict requirements. This classification system is problematic for several reasons: