Social media in the battle for public security
12 September 2011
By Joanne Taylor
We have witnessed a total revolution in the use of social media during 2011. It is no longer just reflecting social attitudes but now defining them. Social media sites have evolved from being a form of communication through to a social activity in its own right and now into a channel for active group psychology so powerful it can over throw governments.
In the world of public security this has massive importance. The Arab Spring witnessed in the first six months of the year proved for the first time that the attitudes expressed through social media are not just the muttering of the young or rantings of the disillusioned, but can reflect the stirrings of a nation or even actually stir the nation into action. Would the events in Tunisia, Egypt or Bahrain have occurred without the presence of Facebook and Twitter, sites so uncontrollable by governments? I suspect not. If undemocratic governments had in fact realised the power of such sites they no doubt would have tried to ensure more control over them as we see in china. But can any government really expect to control the internet, maybe for a while, but in the long term it's virus like nature will no doubt push through the barriers of any government attempts at control.
In the western world we have seen the worst riots for 20 years, London was truly burning and other cities quickly followed. It was not so much social media being used to insight the violence that was new here, but rather the rioters use of social media to co-ordinate their activities. Blackberry messenger became the rioters most useful weapon against the police, being able to target, move and attack en mass, leaving the police trailing behind for three nights in a row. The response of last resort, blackberry offering to shut down its messenger site entirely, if only temporarily.
In the attempt to maintain law and order, are draconian interventions that stop access to social media channels entirely, really the only answer? The answer is emphatically no. I for one would like to keep my civil liberties.
Public security professionals need to stop seeing social media as the problem and start tapping into it as part of the solution. In the fight against anti-social criminals, organised crime and terrorists alike, social media can be the law enforcement professionals most powerful weapon.
Seeing social media as part of the solution, not the problem
Police and Intelligence services across the world have been using material on the internet extensively for a while now in the fight against crime and terror. 'Open source intelligence' the use of publicly available information sources, is fundamental to most agencies daily activities and has proved hugely useful in unlocking a multitude of investigations. For social media the focus has mainly been on what is termed the 'black web' - monitoring 'known' sites that are utilised by criminals to communicate and educate one another, sites used to insight hatred and feed propaganda, or chat rooms used by predators to target the young and na´ve. At an individual level, they use the web to track a person's activity, social network and movements.
As we saw in the terrible events of the July massacre in Norway, Anders Behring Breivik was an avid social media user. He utilised the web to propagate his abhorrent beliefs and gave very clear indications as to his intended actions. Hindsight is a wonderful thing, but the indicators were there, he was known to authorities as a potential threat and was active in extremist circles. So why wasn't he being monitored?
The painful truth is that he is only one of thousands perhaps millions of individuals across the globe that pose a potential threat to the worlds law abiding citizens. So how are public security professionals expected to monitor them all and deal with the massive volume of open source data available through social media, to extract the golden nuggets that will prevent another tragedy.
How do you find the needle in the Social Media haystack?
The argument most public security professionals often raise is that they simply do not have the resources to turn the contents of social media communications into usable intelligence they can act upon. If we are to rely on people as the interpreters of the open source, that is most certainly true. People are not the answer here, technology is.
So why are they not deploying technology to help?
The first challenge for public security professionals is access. Most governments have powers and means to obtain back door access to sites across the internet. The very nature of the web means this is rarely where the issue lies anyway.
The second challenge is knowing where to look. Unlike marketeers who have been utilising social media to better understand how the Public perceives their brand, law enforcement has a much broader set of potential sites and channels that criminals can be using to their advantage. More importantly, criminals are trying to hide, trying to subvert those monitoring and trying to leave as little trail as possible. Because of the sheer breadth of the channels used, and number of individuals of potential interest to law enforcement, the volume of data needing analyses instantly is massive, requiring an industrial scale approach.
The third challenge is the nature of social media, its substance is in the text. Without tools to cut through text at an industrial level and understand meaning from it, we are back to relying on humans reading, tagging and interpreting, to turn data into usable intelligence. However, new innovative technology is available today to exploit the text based unstructured data of the social media.
SAS can provide the industrial scale solution required
Monitoring public conversations across social media requires heavy lifting from text analytics tools, and therefore understanding the technical aspects of text analytics should be a necessity for modern Public Security professionals. Unfortunately it currently is not, in fact the concept of text analytics is often understood little more than the collation of open source data. In fact many companies want professionals to believe that their ability to collate data from the open source (including social media) and simply making it available to run some reports off, is the answer. It is not.
Moving lots of text based data from the web into data stores does not in any way address the challenge of exploiting the full potential of that data and turning it into actionable intelligence. The reason so few companies are really addressing this challenge to help law enforcement and security professionals, is their inability to interpret text based data on an industrial scale.
Some text analytics tools do not just gather data and text, but pore through huge amounts of social media data to uncover patterns and analyze content. Social media analytics can continuously monitor online and social conversation data to identify important topics and content categories to enable professionals to focus in on the content areas they are interested in. Crucially it can automatically extract 'entities' – people, places, locations etc and build linkages to understand the inter relationships as well as context. Social network analysis can be used as part of this process to understand the human networks behind the social media, how people are connected, the closeness of those connections and crucially the ring leaders involved. This enable resources to identify and focus on key individuals to target and so utilise resources most effectively. By assessing and monitoring the 'sentiment' of text it is possible to flag a change of attitudes, that may express that a target individual or group are potentially moving from words to action. Again enabling the technology to do the monitoring freeing resources to take intervene when increased threat is identified.
This type of capability will enable the professionals to filter out all the 'noise' within social media and focus on the data that could provide valuable intelligence. In addition, some estimates are that less than half of all Twitter traffic is in English. Social media analytics can analyze content written in any major language.
There simply aren't enough resources to focus on all negative conversations no matter how hateful or all people known to associate with anti-social groups, but by utilising predictive analytics which groups or individuals that are most likely to pose a threat today or are likely to tomorrow can be identified and flagged. By utilising sentiment analysis combined with advanced risk modeling techniques, those that pose the most potential threat can be identified. This means valuable human resources can focus on where the greatest threat lies and let the technology do the monitoring of the wider community of interest.
Do we trust technology enough to put our security in it's hands?
Hindsight is easy I know, but it does prove the point.
Having reviewed the available news feeds from the Middle East from the first few months of this year, we have shown that text analytics clearly identified the crescendo building to revolution in Tunisia and the same patterns in content categorization and sentiment analysis could have predicted events in Egypt.
Running text analytics on millions of Tweets during the London riots flagged specific locations as targets and groups as ring leaders in real time. Enough to at least keep up with the rioters if not to get ahead of them.
Fundamentally this is about interpreting the text and risk modeling it. Banks, insurance companies and governments have been using this technology for years to identify whether to give you a loan, whether your credit card transaction is fraudulent and whether to you are likely to evade tax. We have to trust it to monitor social media and start to identify potential threats and predict potential wrong doing. In times of financial austerity public security agencies cannot just increase the number of resources to ensure social media is being appropriately monitored for threats. They need to use those resources smarter and free them to action intelligence, not be bogged down sifting through data trying to identify intelligence within it. They need to use technology to fill the gap to extract actionable intelligence from mass social media data.
That means we have to let technology do the watching and trust it to identify threats and flag for human action when needed. An uncomfortable thought for many I know, but I for one would rather a machine is watching than nobody be watching at all.Joanne Taylor is director, public security at SAS