The report describes a fundamental policy problem: the #EvidenceDilemma.

If we wait until risks are conclusively proven, the damage may already have occurred.

But if we regulate too early, we may introduce measures that later prove unnecessary or ineffective.

internationalaisafetyreport.org