Organisations that develop, deploy or use AI systems and any national laws that regulate such use shall adopt design regimes and standards ensuring high safety and reliability of AI systems on one hand while limiting the exposure of developers and deployers on the other hand.