The growth of social media use has made it a vital tool in examining individual and group behaviour. Demos’ #Intelligence report, released yesterday, examines how the police and security services have in recent years turned to social networking sites in their efforts to tackle crime, from terrorist offences to riots.

The task of turning social networking data into usable intelligence – or ‘SOCMINT’ – is daunting, as the figures demonstrate: every day, over 250 million tweets are sent, as are 1 billion messages via Facebook’s instant messaging service. Whilst turning the web into a user-friendly resource for those keeping the peace is far from simple, steps have been made to address this problem and a lucrative online surveillance industry has developed.

Existing technological approaches to internet surveillance rely on powerful ‘trawling’ tools which employ algorithms and other mechanisms to build up a picture of the structure of online communities, the content of online communications and ‘sentiment analysis’ software which claims to detect the tone of texts and audio clips. But these tools suffer from significant weaknesses and sceptical voices question the ability of even advanced technologies to do all that their creators claim. Their sampling methods have been criticised and the demographic differences between the users of websites like Twitter and the wider population (Twitter users are younger, wealthier, more urban and better educated) call into question whether information gleaned from such websites can be accurately applied to the offline world. Similarly, can ‘sentiment analysis’ tools really take things like context and the nuances of online language into account? And to what extent do phenomena such as the ‘online disinhibition effect’ mean that the rantings of an online malcontent can in fact be used as evidence of how they are likely to actually behave?

Using social media sites as intelligence-gathering tools also raises issues, which, whilst not unique to the internet, are certainly heightened in the online world. Misinformation, whether in the form of parodies or deliberate un-truths, is especially prone to be spread through social media channels – during the riots in London, Twitter acted as a conduit for a slew of rumours ranging from warnings of freed tigers roaming Primrose Hill to reports of tanks deployed at the Bank England.

Another problem is ‘gaming’ – whereby aliases are assumed and ‘facts’ reported pseudonymously – as the case of the unmasked blogger behind A Gay Girl in Damascus demonstrates. The growth of online identity theft heightens this risk, as opponents of online surveillance measures are quick to point out: how can those carrying out the surveillance really be sure that the people behind online identities are who they say they are? If the gathering and use of SOCMINT is to be effective and legitimate, the public must have confidence that the police and security services are capable of counteracting these threats.

It is in light of these risks that the report calls for increased investment in the technological and human capabilities in the sphere of SOCMINT. First, the development of computing tools capable of identifying data of intelligence value whilst also taking into account the problems of context, online culture, gaming and reliability is absolutely crucial. Second, training people how to operate these tools and understand how to collect, identify and analyse SOCMINT is equally vital – one of the most interesting insights to emerge from the HMIC report into policing during the August riots was that officers lacked the basic competencies in this area, a weakness which clearly must be rectified.

These two conditions are necessary if SOCMINT is to be integrated into existing intelligence structures, as the report advises. The rise of social media has been one of the most notable features of the past decade and operational guidelines and legislation are often worryingly out of touch. For law enforcement agencies, engaging with online communities is no longer a choice but a necessity and in order for them to fulfil their duty to protect the public a revised approach to SOCMINT is needed; one which fully values its importance whilst simultaneously recognising its potential weaknesses.

New Comment