From OSINT to SOCINT. How social media is turning into a battlefield between intelligence agencies and society

This is the request for proposal through which the FBI is making public its intention to purchase a program for monitoring social media, with the stated objective of exploiting social communication means in order to detect and investigate a diverse gamma of threat towards US national interests: terrorist groups, internal threats, organized crime activities.

In the meanwhile, in Moscow, hackers managed to steal 7.5 TB of data, some of them referring to secret projects developed by the FSB to control social networks.

New concepts and areas for a more thorough Intelligence.

The face that secret services obtain their data also from virtual and non-virtual mass media is not a secret, with the concept of open-source intelligence (OSINT) already being a familiar one: gathering data from opens sources… any data which can legally be obtained from public sources. Technically, OSINT also houses any public information, such as articles, press releases, books or documents in a public library, but in practice it especially refers to information found on the internet.

The information is gathered from public sources by secret services, but also by companies and organizations, which need data mostly to draft defensive strategies, by knowing the potential adversary’s characteristics. Their collection is facilitated by different intelligence instruments, programs which can perform complex scan, from gathering email addresses to social media or dark web sources.

Recently, the concept of social media monitoring was added to OSINT, a term presented in the FBI’s request for proposal and which, at a first glance, is not that hard to define or understand.

However, it can be seen most of time without rigorous delimitations, together with social media listening, social media analytics or social media intelligence. Or with online analytics, buzz analysis and social media measurement.

For some, these phrases are most synonymous, while for others their name realities with subtle differences between them.

For example, social media listening would show information on media consumers, concentrates on gathering knowledge of the public and emotions, while social media monitoring would only aim to maintain a reference point, a standard obtained in relation to the target public through social media listening. Social media monitoring would be a passive database of social media data, while social media listening presumes actively listening to what is happening in the environment and taking actions based on it, which warrants a deeper analysis of the term social media monitoring, regarding what is observed and understood in an initial phase.

At the same time, social media intelligence (SOCMINT) refers to the technological means used to monitor social environments, including social conversations, and later to analyse and use the significant content in decision making. The term was introduced in a 2012 paper authored by David Omand, Jamie Bartlett and Carl Miller, who considered it as the most recent member of the intelligence family which includes IMINT, HUMINT, SIGINT, OSINT etc.

And because things were not sufficiently complicated at a conceptual level, the term SOCINT (Social Intelligence) is also passed around, without a unitary definition. Laura K. Donohue (The Dawn of Social Intelligence) uses it while specifying that it is a potential member of the HUMINT, SIGINT, OSINT etc. collection which refers to digital data regarding social relations, the extraction from digital sources (metadata, social mass media, geolocation info) of data destined to build a detailed picture of the networks. However, for Kerry Patton, SOCINT is a key-instrument in intelligence operations, presuming the observation and analysis of individuals and communities, understanding the behavioural aspects of relations and systems, with attitudes and social cultures integrated in the analysis of information collected to maximize the efficiency of security initiatives.

Social media – a valuable data base with apparently unrestricted access

Essentially, social media monitoring is the act of gathering online data (from social platforms) and content (forums, blogs), which need to be analysed in order to draft a strategy, based on understanding the social phenomena and evolutions in the targeted communities.

Observing the competition’s online posts, but also those in its own business area, those made by targeted groups, new tendencies, consumer problems and many others is the best way to identify the opportunities which a certain mark of product can benefit from. Companies are using data from social media to understand their client base better and to efficiently manage marketing efforts. After the data is gathered, an analysis on it is performed to understand the public and groups of people. Therefore, monitoring social channels and online conversations is the first step. Analysing the gathered data, differentiating between authentic and fake news is the next step towards drafting a strategy and creating a marketing campaign based on a detailed analysis of the audience’s emotions, opinions and behaviour.

This also applies to potential adversaries. Because social media offers and immense quantity of information on individuals – their political and religious preferences and more, the identity of friends and family – it was proven to be irresistible both for commercial entities, as well as governments, security agencies, defence and public order.

Naval transport companies monitor the Somali pirates’ Twitter and Facebook messages, in order to understand the manner in which criminal organizations are targeting the ships and planning the attacks. A geographical post made by an ISIS fighter on a social network can be useful for an intelligence agency.

The analysis of data obtained from social media has a significant potential to support the operations of intelligence communities, offering an image of the ideas, intentions and opinions of relevant segments of the public. This can help, for example, with identifying extremists and groups undergoing radicalization. Last, but not least, the details obtained through this analysis could also be used to influence perceptions, behaviour, decisions.

Technological solutions to analyse social media information

Electronic communication has increased the capability to rapidly transfer and use large quantity of data, but has also created new vulnerabilities. The large volume and diversity of social media information – Twitter users alone can post up to 500 million tweets per day, photographs, clips and status updates, profiles which include personal details many times – creates problems in selecting the relevant information necessary in planning and carrying out civilian-military operations destined to accomplish new defence and security objectives.

Processing this huge quantity of data calls for specific instruments, programs destined for the real time collection of data on threats and security events, algorithms which will allow to understand the activity of a criminal or terrorist group, contribute to its identification, and also understand public opinion or modify it.

Data gathering, but especially selecting and extracting the relevant information, needs special software, web crawlers, which make a first filter of selections and reducing the quantity of information which can be later made available analysts.

Social media monitoring has a significant number of informatic instruments which differs with regards to the accuracy and speed with which data is selected, the capability to use several languages, including slang and emoji, to analyse the subjective opinions behind the posts, the capacity to integrating. We are talking about software which can analyse web applications, of which most function by continuously accessing and indexing websites.

Maybe more known to the public are Google Analytics, Twitter Analytics, Facebook Insight. According to a 2018 Electronic Privacy Information Center reports, suggestively titled Government Surveillance of Public Space, there is a significant number of instruments destined for social media monitoring, a number which has been significantly on the rise in the past couple of years. In the US, more than 500 companies, politicians, federal agencies, intelligence agencies, defence contractors and the armed forces have purchased such products: XI Social Discovery, Geofeedia, Dataminr, Duanmi, MediaSonar, SocioSpyder.

The CIA even has an investment fund, In-Q-Tel, established in 1999, in order to identify innovative technology solution to support the agency’s mission. According to the presentation on its own website, “from the dot-com boom up until the era of artificial intelligence, In-Q-Tel” was interested in revolutionary instruments of technology “in order to help intelligence and defence communities keep the nation safe”. Next to social media investments, In-Q-Tel developed a special technology lab in Silicon Valley, named Lab41, where intelligence and technology industry specialists are concerned with obtaining a better understanding on how to work with – and eventually use – data analysis.

In 2004, the FBI also became an In-Q-Tel partner, followed by the Defence Intelligence Agency in 2005, the NSA in 2009 and the Department of Homeland Security in 2010.

In-Q-Tel has invested, among others, in Visible Technologies and Palantir Technologies, two important providers of instruments meant for social media monitoring. The Visible Technology software monitors Twitter, Facebook, Flickr, YouTube, Amazon and is capable to monitor communications in real time and provide analyses. Palantir Technologies, a start-up owned by PayPal co-founder Peter Thiel, who is also a member of the Facebook board of directors, is famous for reportedly helping, with its technology, at the capturing of Osama bin Laden.

At the same time, the FBI has a contract to use the algorithm provided by Dataminr, which is capable of monitoring Twitter, and purchased SocioSpyder, an investigation software which gathers data from several social media platforms.

In the meantime, intelligence agencies have become a part of the phenomenon, with their own Facebook and Twitter accounts and use social media both as a source of data gathering as well as a manner to present themselves and interact with the public opinion.

The exploitation of social media – between legality and the right to confidentiality

The monitorization of social networks is legal, but even though the information is technically open source, and the creators of monitoring instruments claim that they act only on what is publicly available, the use of data obtained in this manner for internal investigations or operations by the agencies can be problematic.

Critics claim, for example, that in the lack of minimal surveillance, intelligence agency employees could be tempted to use social media monitoring tools to compile information on politicians, journalists and other individuals and exploit them for political gain. Moreover, they also claim that social media monitoring algorithms can be discriminatory and have high error rates. Lee Rowland, a lawyer in the American Civil Liberties Union, states that there is a dangerous tendency from the government, based on technology companies, “to construct massive files on individuals”.

Privacy International, a British charity society based in London, which defends and promotes the right to privacy, suggests that it would be necessary for the term SOCMINT – defined as the analytical exploitation of social media information – to be used as a different concept from OSINT, arguing that the surveillance of social communications means should be eliminated from the definition of OSINT. This would ensure a more nuance regulation of the area, with protection measures and policies which will taking into account the unique and specific properties of social media, of a space which is difficult to categorize into public and private.

Intelligence agencies should gather, use, transmit and keep social media data only with adequate knowledge and use of internal and international law elements, by respecting ethical principles and applying some measures destined to respect confidentiality.

And not only by the intelligence communities, or by the US. Facebook finds out all your food preferences. Uber knows where you usually travel. In China, until 2020, the entire population will be registered in a large national database and each citizen will have a social credit score, so that depending on it, an employer can asses his future employee, or the landlord his future tenant.

In a world described by Google Director Eric Schmidt as producing within the span of two days more information than it did from the beginning of civilization to the present, it is necessary to keep a reasonable balance between the need to ensure security and to maintain privacy, protect the private life and personal rights, as well as an answer to the public’s expectations of transparency.

Translated by Ionut Preda