In the case of NCII, I noted earlier this year that tech firms seemed to want to pretend that any false positive is "not actually a false positive", because it "looked close enough" by changing the definition of NCII to include the false positives. It's an absurd attempt to evade scrutiny of their moderative practices.