If AI flag and human is "immaterial" difference as you noted, wouldn't AI flagging it as abuse material make it impossible to view it by a (non-LEO) human after

If AI flag and human is "immaterial" difference as you noted, wouldn't AI flagging it as abuse material make it impossible to view it by a (non-LEO) human afterwards, as now that would show actual intent to view that material (it is illegal to view it with intent to view it). Surely your reviewers never deliberately looked at the material a second time after marking it as abuse material for law enforcement, as that would be legally risky I would think. The issue here is even if local LEO clear it, the federal prosecutor's opinion can differ and he can use the fact you knew AI flagged it and that you released it anyway to damn you. There are cases out there where local/state government cleared someone of wrong-doing but they still got hemmed up federally.