Why not be angry at all of them?

As someone who works with sensitive healthcare data, I can tell you that the mere existence of a dataset doesn't guarantee its misuse, nor it does it absolve anyone who interacts with that data of responsibility for proper stewardship.

Yes, you are right that we should think carefully before creating a sensitive dataset. If we insist on creating such a dataset, the people involved must put in place guardrails for stewardship of those datasets. But the stewards of that data, past, present, and future, also share responsibility.

Of course if the incentive structures don't line up with concern for mitigation of harm to vulnerable people as is the case with law enforcement in the US, then all of that is out the window.

Anyway, what you have written implies that we need not think about accountability for those who misuse of datasets after they are created, which is clearly absurd as I and anyone else familiar with healthcare data can tell you.