Minority Report – Predicting a riot is impossible on social media

In the wake of the UK-wide riots and the growing evidence that mobile applications like the BlackBerry Messaging system had helped rioters organise and amplify their activity, the UK government is considering a review of social media. Specifically, Home Secretary Theresa May plans to sit down with Twitter, Facebook and RIM, to discuss the issues. That’ll be an interesting chat.

She hasn’t minced her words: “Social networking sites like Facebook and Twitter and messaging services like Blackberry Messenger have been used to coordinate criminality, and stay one step ahead of the police… I will convene a meeting with ACPO, the police and representatives from the social media industries to work out how we can improve the technological and related legal capability of the police.”

She also said: “Among the issues we will discuss is whether we should disrupt messaging services when trouble is being planned.”

Her statements followed Primeminister David Cameron, who told parliament yesterday that “instant messaging services” (as he put it) will be reviewed. “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,” he said.

Now, to my mind there is a lot of confusion going on here. Private messaging services like BBM are being confused with public social media services like Twitter – the latter of which helped the post-riots clean up enormously, and the chatter on which often prevented people from wandering into dangerous zones. It’s even helped an old man get his small businesse get back on it’s feet. Some Police forces positively love it.

The media (and the tech community) has leapt upon the Government’s statements as indicating that they plan to ban a) social platforms b) ban the people who incite riots from social platforms.

Of course, the government review will come up with obvious answers. You can’t ban social networks and you can’t ban people from using them – they will find a way.

More importantly, what the above statements indicate is that the government would like some kind of early warning system for the Police, to alert it when people are “plotting violence” or when “when trouble is being planned” on social networks.

Facebook has already reacted by saying it’s taken several potentially dubious groups off its network. Twitter has said it’s “happy to talk”. RIM is “cooperating”.

But as any keen observer of social media will tell you, trying to track sentiment on these platforms is very hard. Indeed, this problem vexes Facebook and Twitter, who have millions in funding, all the time. Do we really think the UK government is going to be able to crack that one?

I asked Nick Halstead, CEO of DataSift, what he thinks. Halstead has been doing battle with MP Louise Mensch in this very issue.

He told me:

“Just as Hedge funds are already looking to Twitter to build insights into stock movements, it is just as plausible to intelligently describe a certain behaviour and pro-actively look for patterns that may relate to unrest. Twitter themselves already detect trends but these are calculated from a global or country perspective. It is possible to apply the same trend detection techniques within a more highly curated stream. And those streams can be defined by looking at the topic of conversation, the intent of the message and sometimes even the location.”

So in other words, you could potentially track something. But would it be accurate enough?

Could you predict any kind of incitement to riot on social networks?

“I think that’s far too esoteric a concept right now for computers to understand,” he told me.

And there’s the rub. One person’s BBM or Tweet saying “It’s all kicking off in the high street” could be interpreted as an incitement to others to join a riot – or as a warning to their friends. And let’s not even go into how an algorithm is going to decipher the language of the street.

Plus, the Home Secretary is going to have to do more than visit Facebook, Twitter and RIM to deal with the issues of the spreading of messages about riots.

Our research has found healthy use of the WhatsApp application among London’s youth (although BBM is far and away preferred). And let’s not go into Kik, PingChat and the rest of the group messaging startups.

And there’s the small matter of Facebook Messenger coming out, and Apple’s iMessage. All of these will be group messaging leviathans.

Put simply, the private mass messaging genie is out of the bottle.

So the irony of the situation is that in order to get an early wind of a “situation”, the government should be positively encouraging people to join social networks, not come off them or shut them down.

Apart from anything else, the more data there is in the system, the more the crowd can deal with people who might be inciting a riot and bring them to the awareness of the authorities.

Indeed, the Police should not be looking for ways to shut down the BlackBerry BBM (I’m simply going to assume they will ask GCHQ to snoop on it anyway, and I’d be amazed if it’s not already).

Instead, the Police should simply grab the ecosystem with both hands. The Guardian and other media showed how it was child’s-play to simply get a Curve handset, circulate a PIN and then get forwarded BBM Broadcasts about riots and looting by concerned citizens.

Hey, how about even some old fashioned Policing, whereby you get an informant to forward the dodgy messages?

But the fact that the @metpoliceuk Twitter account, has made 589 tweets in two years and remains a broadcast-only account with zero interaction with the public, speaks volumes about the Police’s knowledge of how these platforms work.

So let’s fix THAT problem before we talking about monitoring.

Far better for the Police to set up a special social engagement unit – or something similar – than for us to end up with a Minority Report style society.