The Group of Chief Scientific Advisors recommends that, before widespread AI deployment in EU crisis management, several critical conditions must be established, including formal risk assessments focusing on cybersecurity, infrastructure resilience, and strategic autonomy, particularly regarding non-EU AI systems, as well as an assessment of the social acceptability of these risks. They also recommend the systematic consideration of the opportunity cost of major AI deployments, and make a plea for better data preparedness through harmonised collection standards, secure sharing protocols, and resilient infrastructure with analogue fall-backs. They stress the need to keep humans at the centre of all decision-making, as crisis decisions carry profound moral and social weight that AI cannot appropriately handle. The overarching principle is that AI development must complement, not replace, essential improvements in European governance, monitoring systems, and crisis prevention strategies.