Comment by __MatrixMan__

Is this because your threat model doesn't include your device vendor/admin behaving badly? Or because you understand something about the protocol which prevents them from doing so?