David Cameron has vowed to do ‘whatever it takes‘ to restore order after the riots: this seems to include blocking access to social networking sites.

“Everyone watching these horrific actions will be struck by how they were organized via social media,” Cameron said. “And when people are using social media for violence we need to stop them. So we are working with the police, the intelligence services and industry to look at whether it would be right to stop people communicating via these websites and services when we know they are plotting violence, disorder and criminality.”

We’ve heard a great deal about the use of social media for rioters to organize, as well as boasting about their ill-gotten gains – posting images of stolen goods and inciting others to do likewise. Exactly how much incitement to riot was occurring in these spaces is unclear. Mobiles, too, are implicated, as BlackBerry’s BBM encrypted messaging service was purportedly used by some to coordinate looting.

Is it desirable, or even possible, to shut down access to social and mobile media networks during times of crisis? Exactly what does Cameron have in mind?

To find examples of technical feasibility, we have multiple examples in the Middle East, where access to the internet has been cut off several times during the events of the Arab Spring. Egypt, Libya and Syria have all severed net traffic during protests, and with considerable effectiveness. It’s not the kind of list we’d expect a mature democracy like England to want to join. If Cameron is thinking about something more individualized, where only some people are restricted, this will still land him with some very sticky civil liberties issues.

But is shutting down networks – in part or whole – even a good idea? To answer that seriously, we need to look at the whole range of activities that occur in social media spaces like Twitter and Facebook. Social media spaces are complex: they perform critical roles of sharing information, news gathering and emotional support, while also being conduits for rumor and abuse.

If we take the case of the London riots, we have seen extraordinary acts of pro-active social engagement: from #riotcleanup which rallied people to clean up the streets, to various forms of citizen journalism that have made substantive contributions to media coverage (for example, see Leon Piers‘ @BristolRiots on Twitter or Casey Rain’s Tumblr, Birmingham Riots 2011). Malaysian student Asraf Haziq, who was attacked at knifepoint by rioters, became the focus of a social media campaign to raise money for his recovery.

These significant social responses, organized and distributed via Twitter, Facebook and Tumblr, would cease to function if networks were restricted or shut down. We can consider the people on social media sites who engage in community rebuilding as ‘first responders’: before government programs respond or insurance claims can be filed, people are organizing via a range of space (some online, some not) to organize and assist those who have been most affected by a disaster or a riot. We saw similar actions in response to the devastating Queensland floods: first responders used Twitter to share information and assist flood-affected communities. If governments restrict or paralyze these networks, they also bring to a halt precisely the kinds of pro-social organizing that communities use to fill the gaps long before many official institutions can respond effectively.

Instead of considering the complex ecology of social networks, we are being told they need to be shut down and, as Cameron argues, greater policing of communication channels. There’s considerable research that points to this being ineffective, and more likely to increase community fragmentation and anger (see, for example, the discussion of the sociological literature here) There’s a need for more research to show the complexity of activities that occur in network spaces during acute events (I’ve been doing some work on this in Australia with Jean Burgess and Axel Bruns). But while many of the negatives (looting, rioting and the like) will occur regardless of social media’s availability – we need only look to Paris in 2005 or countless other ‘pre-Twitter’ riots – many of the ‘first responder’ positives rely on social media networks to function. The #riotcleanup idea could not have occurred on such a scale had it not been for Twitter. As Debbie Chachra writes on Twitter: “Urban rioting existed before SMS/social media. You know what didn’t? Large-scale community cleanups, spontaneously organized within hours.”

If the riots remind us that cities are always vulnerable spaces of politics, negotiation and rupture, Cameron’s realpolitik underscores that social media networks are also fragile. They are susceptible to challenges, both technological and political: including hackers, accidental system overloads, and politicians who wish to censor, restrict or block them, be they in democracies or dictatorships.