My immediate thought would be to focus on the mobile space first, rather than the workstation-based internet platforms, like laptop or desktop computers. The mobile space is growing faster and is much more pervasive in many of the at-risk areas where I would imagine you are focused. From a technical, implementation standpoint, companies like Clickatell already provide the capability to reach mobile users across different carries in countries like Afghanistan and Pakistan with a relatively simple, commercial platform and API. A strong presence in the mobile space will create room to then drive people towards more content-rich and flexible platforms, like youtube or facebook, which could carry an even stronger message. Some new media platforms, like twitter, coexist within both spaces as well. I apologize for not having any case studies or examples at hand, but I thought this bit of information might be useful from the tech/implementation side.

Another consideration when looking at social media in a counterinsurgency capacity is to anticipate and plan for the intimidation that will likely arise when insurgents begin to target members of the population who are identified as using these technologies, whether mobile phones or computers. Any new media strategy must be built upon the ability to protect the population (or potential terrorist/insurgent recruits that we wish to influence) while also providing the communications channels for effective IO. Just a thought.

Hopefully this will at least give you a few points to think about!