The thing that surprises me about the coverage of Apple's new child abuse photo detection thingy is that it appears to me that it might be counterproductive from the POV of protecting children against abuse. The mechanism is intended to detect _known_ child abuse photos and is publicized. If I wanted to have access to child abuse photos, knowing about this mechanism would make me try to acquire photos that are not well known, or that are very recently taken. So, assuming rational behaviour, this shifts the demand from photos of child abuse to new and unpopular photos of child abuse, which increases the amount of child abuse happening (assuming there's some market with enough information sharing to have sane supply/demand dynamics).

Is this reasoning incorrect somewhere? Is everyone (else) practicing the virtue of silence on this topic, given that the effect is magnified by public discussion thereof?