Regardless of the means, any entity which performs some form of censorship --- especially if that entity is through which a nontrivial amount of the population gets their knowledge and can be influenced by, like Google --- automatically makes itself the subject of moral and ethical scrutiny.
There's no reason to think a human was even involved in the decision to apply this label.
The decision was made long ago when Google started to classify sites as "unsafe" or "safe", and that was definitely a human decision. IMHO Google should still be held accountable for whatever effects their automated systems have.
This reminds me of how cracks, keygens, and other... anti-authoritarian (for lack of a better term) tools will often be detected by antivirus as "unwanted" and/or "suspicious", despite not actually doing anything the user didn't want. It might just be heuristics that happen to cast such a wide net, but it also means the authors made that decision, that moral and ethical choice.
The site may have been hacked (and just because it looks OK when you visit doesn't mean it hasn't been hacked). It could even be a state-sponsored attacker trying to exploit a zero day to monitor wikileaks visitors. Or, as someone points out below, it could be run-of-the-mill malware contained in attachments of some of the recently released email dumps.