In the proposal for the new AI regulation, the EU Commission formulates risk-based, horizontal requirements for AI systems, supplemented by a voluntary code-of-conduct for AI systems that do not represent a high risk, and a "regulatory" sandbox to promote innovation.
