Danger, harm and change
Just because a place is currently "safe" for research sources and collaborations, it does not mean it will always be safe. The potential harms to those cooperating on research projects has been well-outlined in some of these processes and in the article collections related to this project. What has been said less is that, well, any state can develop authoritarian tendencies, and the discussions of post-election US indicate this. Nor does a state does not have to be fully authoritarian to place sources and collaboraters in a condition of harm. I have had arguments at confernces with advocates of transparency over this. In their view "it will never happen" that sources could be harmed. Or, "it will never happen" that qualitative research in touchy places will be disadvantaged, because apparently journal editors are always wise enough to know which pieces should be exempt from submitting data. But what was ok last year for Turkey is no longer ok. Even the sagest journal editor would not have forseen this.
There is a real need for strong, institutional / formal protections against potential harm for sources. And in fact, those protections, to be robust, should not be granted only to certain countries, but rather granted with an eye to the fact that political systems can become repressive while a piece is under review, or materials submitted to a reviewer could be accessed years later and used in ways that harm sources an collaborators. It is not acceptable to be so complacent about the potential harm to sources as the current advocates of DA-RT have shown to be.