Social media analytics: Security and privacy issues

Social media (SM) have gained widespread usage in public and private domains. Individuals and organizations use SM to express themselves, to shape political agenda, to gain public acceptance, to build organizational image, and to distribute propaganda, among many possible uses. As usage of SM increases, issues concerning security and privacy also arise. Examples abound. Terrorists and extremists frequently distribute their propaganda and recruit their new blood strategically on SM (Chen et al., 2008). Hackers constantly search for exploits and weaknesses in SM platforms to disrupt government and business operations. The recent Twitter-AP attack shows vulnerabilities in these platforms (Selyukh, 2013). Cybercriminals steal and abuse private information available on SM forums, smartphone applications, and online accounts to gain illegal access. The recent hack of the U.S. Office of Personnel Management by a teenage-hacker exposed the risk. The hacker broke into the AOL account of Central Intelligence Agency (CIA) Director John Brennan and stole sensitive documents, including his application file for top-secret government security clearance (Messing, Schram, & Golding, 2015). Meanwhile, states and non-state entities pose threats too, as seen in cyber attacks against Estonia and eastern Ukraine (Lange-Ionatamishvili & Svetoka, 2015). In the business world, companies and hacktivists are increasingly mining for private data from SM and applications. Many social networking applications allow massive collection of fine-grained data that is tagged with rich personal details (e.g., demographic data), geographic location (e.g., IP addresses), and temporal usage (e.g., user access log data). As an example, publicly available data and some analysis efforts could reveal personal identities and their taxi tipping habits (http://chriswhong.com/open-data/foil_nyc_taxi; Trotter, 2014). If used improperly, these applications could pose significant threats against user privacy and systems security.

Social media analytics (SMA) is a set of informatics tools and frameworks to collect, monitor, analyze, summarize, and visualize SM data, to facilitate interactions, and to extract useful patterns and intelligence (Zeng, Chen, Lusch, & Li, 2010). SMA has been applied to business and public domains, such as crime analysis, marketing, and public relations (Fan & Gordon, 2014). Its key capabilities include universal reachability, speed of interaction, multi-modal communication, device compatibility, and emotional appeals of SM. These capabilities can be used to enhance security and privacy protection. In the wake of online security and privacy problems, there are urgent needs to develop new approaches to SM-based cybersecurity informatics and to understand the SM privacy and security issues. In this Special Issue of JIPS, three research articles and one book review address different yet important aspects of these needs.

The first research article titled “A simulation-based approach to predicting influence in social media communities: A case of U.S. border security” describes a new approach to simulating activities of SM communities on U.S. border security discussion. The author developed and applied the approach to analyzing millions of tweets posted by hundreds of thousands of users. Based on a comprehensive grid-search and a rigorous evaluation, the predictive models were shown to accurately model real-world SM community behavior. The article also offers recommendations for further research in modeling and predicting human behavior in SM communities.

The second research article titled “The role of personalized services and control: An empirical evaluation of privacy calculus and technology acceptance model in the mobile context” examines user privacy perception on social-networking services delivered through smart phones. Based on their proposed framework combining privacy calculus and technology acceptance model (TAM), the authors conducted an experiment of various personalization settings of a geographical-location-based social network of restaurant services with 308 subjects. The experimental results show that when the perceived benefit is high, users tend to be more willing to disclose their personal data in a mobile social-media-based networking application.

The third research article titled “Personnel security and open source intelligence: Employing social media analytics in pre-employment screening and selection” focuses on security checking of personnel recruitment using SM. The authors reviewed background investigation manuals from U.S. state peace officer accreditation organizations and found no standardized protocol for pre-employment personnel screening and selection using open source intelligence and SM. Then the authors developed relevant recommendations based on U.S. federal security clearance process for investigators and personnel administrators.

The book review is titled “Proceedings of The 2015 NSF Workshop on Curricular Development for Computing in Context.” The authors summarized the outcomes of an expert panel discussion that is featured in a published paper (Chan, Chung, & Plante, 2015) of the Proceedings of The 2015 NSF Workshop on Curricular Development for Computing in Context (2015). Based on the authors’ summary, the expert panelists stressed the importance of (1) teaching computing in context by giving relevant examples, (2) deeper learning to enhance career development, (3) recruitment and retainment of women in computing careers, (4) undergraduate research in computer science, and (5) placing students in the context of their culture. This advice is very timely for developing SMA talents and next-generation computing professionals.

Overall, this Special Issue of JIPS provides a comprehensive view of using social media analytics to address emerging and existing needs on security and privacy issues. Decision makers, security administrators, researchers, and students should benefit from the insights and findings from these articles.

Last but not least, I would like to acknowledge the generous support from the JIPS editor-in-chief, Dr. Kallol Bagchi, who provided valuable advice and timely assistance on the preparation of the Special Issue. I also thank the anonymous reviewers for their expert comments and careful review, and the editorial staff and publisher for their assistance.