The Commission should be in a position to take coordinated action or request the market surveillance authority of one Member State to do so where there is evidence that a specific AI system presents a serious risk to health, safety, fundamental rights or the environment, or where the provider operates in several Member States without establishing a legal representative within the Union.