The obligation under the EU AI Act for providers of high-risk AI systems to immediately report any serious incidents (e.g., those leading to death, serious harm to health, or irreversible disruption of critical infrastructure) to the relevant national market surveillance authorities.