is privacy meaningful if it depends on provider honesty?

​

if a system requires trust in a centralized party to maintain privacy, then the privacy itself becomes conditional rather than inherent, which raises the question of whether such a system can truly be described as private in a strict sense. this becomes more complex when users have no direct method of verifying the claims being made about data handling.

so i am wondering whether architectures that remove the ability to observe entirely are closer to the ethical definition of privacy, or if trust-based systems are considered sufficient within practical constraints