Strategic Social Network Analysis

How can individuals and communities protect their privacy against social network analysis tools? How do criminals or terrorists organizations evade detection by such tools? Under which conditions can these tools be made strategy proof? These fundamental questions have attracted little attention in the literature to date, as most social network analysis tools are built around the assumption that individuals or groups in a network do not act strategically to evade such tools. With this in mind, we outline in this paper a new paradigm for social network analysis, whereby the strategic behaviour of network actors is explicitly modeled. Addressing this research challenge has various implications. For instance, it may allow two individuals to keep their relationship secret or private. It may also allow members of an activist group to conceal their membership, or even conceal the existence of their group from authoritarian regimes. Furthermore, it may assist security agencies and counter terrorism units in understanding the strategies that covert organizations use to escape detection, and give rise to new strategy-proof countermeasures. Introduction Many problems in social network analysis (SNA) have received considerable attention in recent years across various disciplines, including Artificial Intelligence and MultiAgent Systems (Sabater and Sierra 2002; Nguyen, Kowalczyk, and Chen 2009). Scientists, developers, and analysts have focused on improving the performance of various SNA tools, such as centrality measures (Koschützki et al. 2005), community-detection algorithms (Orman and Labatut 2009) or link-prediction algorithms (Getoor and Diehl 2005) just to name a few. Unfortunately, while such tools have many legitimate applications, they can be also used to invade privacy or even undermine security of individuals and groups. For instance, by analysing Facebook’s topology as well as the attributes of some users, it is possible to infer attributes of other users (Mislove et al. 2010). Furthermore, such an “attribute inference attack” (Zheleva and Getoor 2009) can be strengthened if it is preceded by a “link prediction attack” which aims at revealing the links that appear to be missing—either deliberately or otherwise—from the topology of the social network. ∗Both first authors contributed equally to this paper. Copyright c © 2017, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved. Naturally, various countermeasures supporting privacy and security that have been proposed including strict legal controls (EU: 2015), algorithmic solutions (Kearns et al. 2016), and market-like mechanisms with which the participants are able to monetize their personal information (Lane et al. 2014). They are, however, difficult to implement in practice, especially at a global scale. For instance, it is highly unlikely that legal privacy-protection mechanisms will be enforced by authoritarian regimes who typically censor internet content including social media (King et al. 2013). Against this background, we ask the following question: How can members of a social network strategically manipulate their online data in an attempt to evade SNA tools? Addressing this question may help members of the general public to better protect their online privacy and security. It may also help activists groups in avoiding censorship. Furthermore, it may assist security agencies and counter-terrorism units in understanding the strategies that covert organizations use to escape detection. In particular, recent findings on covert organizations—especially with respect to the techsavvy ISIS—clearly demonstrate their ability to neutralize counter-terrorism efforts by the authorities. The known evasion techniques used by ISIS range from changing aliases and keeping personal profiles private (Nordrum 2016) to using encrypted communication platforms (such as Telegraf (Khayat 2015)) and staging the disappearance of an entire group from social media only to pop up again in a different place under alternative aliases (Nordrum 2016). In fact, it is believed that the evasion capabilities of ISIS significantly increased after Edward Snowden’s disclosure of classified information on the SNA techniques used by US intelligence (Scarborough 2014). Unfortunately, neither do we have sufficient understanding of such evasion techniques nor do existing SNA tools have the ability to internalize them. This is because most SNA tools were built around the assumption that individuals or groups in a network do not act strategically to evade those tools. Even the more advanced tools that are especially designed for analysing covert networks (Perliger and Pedahzur 2011) typically assume that the network under investigation is not subject to strategic manipulation. Given this, we believe that the literature has now reached a point where serious attention should be directed towards the strategic evasion of SNA tools and building new strategy-proof tools.