That's a very interesting question that presumes that any policy problems at all are among problems that computer scientists should be addressing. IMHO pretty much any solution to a policy problem would by definition not be a computer science problem.

Developing a very specific technology might help solve a social problem if it's applied properly and if social scientists believe that this particular technology will actually have the intended social effect and not backfire.

Other than that, I'd argue that computer scientists can (and likely should) have an informed opinion on various policy problems and how technology could affect them (or be affected by them), but that it's not their role, skillset or responsibility to address any policy problems; if they're tackling policy problems, then they're doing it "wearing a hat" of a citizen-activist, but not really as a practitioner of computer science - if there's an argument that they should be tackling a particular policy problem, then that's only if everyone should be tackling that policy problem, not particularly for the field of computer science and the related technologies.