You're not wrong. Privacy mechanisms still need to be designed and implemented. There is much research, particularly in the landscape of zero-knowledge proofs.