I don't think "can be used to gain access to systems" is a good working definition of "useless".
I think you mean something like "changeable" or "able to be made secret again". If someone finds out you're bribing the police and publishes it, you obviously can't change that fact or somehow directly mitigate the damage. But some privacy issues are more like passwords, they cause damage only in the right situations, and the window can close.
So while there are differences, I don't think there are two neat categories like you suggest.
Doctorow: The Curious Case of Internet Privacy (Technology Review)
Posted Jun 13, 2012 0:49 UTC (Wed) by JanC_ (guest, #34940)
[Link]
It's also not always obvious what the potential damage is or can become at the time the "disclosure" happens (e.g. because information might not be harmful on itself, but when combined with other information it can become dangerous).