After almost three years of political negotiations, the EU has succeeded in adopting a flexible and cooperative digital law. Based on internationally accepted principles and a risk-based approach, the Artificial Intelligence Act focuses on high-risk artificial intelligence (AI) systems and shares the regulatory burden fairly among various market actors. New approaches (i.e. regulatory sandboxes, technical standards) should help SMEs to develop compliant AI systems more cost-effectively.
Nevertheless, we have serious doubts if the product safety approach is conceptually capable to regulate an evolving technology. The often vague and contradictory law with its overcomplicated governance system risks to repeat many problems from the time when the GDPR became applicable. Our AI developers will often not know how to comply with the AI Act and who to turn to if they face problems.
As many voices from civil society, science and industry have nevertheless called for the adoption of the law in recent weeks, we, the CDU delegation in the European Parliament, voted today – despite our concerns – in favour of the AI Act. To ensure that legal uncertainty does not hamper the development and deployment of AI in the EU, Axel Voss has however published a list with 10 additional measures on his website.