The range of malicious activities perpetrated on online social networks is regrettably wide, ranging from malware and spam to controlling fake and compromised accounts, to artificially manipulating the reputation of accounts and pages, spreading false information and terrorist propaganda. Unfortunately, research in this field has looked at these problems in isolation, almost exclusively relying on algorithms aimed at detecting malicious accounts that act similarly. This ultimately yields a catandmouse game, whereby providers attempt to make it more and more costly for fraudsters to evade detection.
This prompts the need for a multifaceted, multidisciplinary, holistic approach to advancing the state of knowledge on cybersafety in online social networks, and the ways in which it can be researched and protected. In this field, there exists a number of interconnected, complex issues that cannot be addressed in isolation, but have to be tackled and countered together. This Dagstuhl Seminar, we identify and plan to focus on the most relevant issues in cybersafety, as well as to explore both current and emerging solutions. Specifically, we identify four problems that are the most pressing both in terms of negative impact and potential danger on individuals and society, as well as challenging open research problems calling for a multidisciplinary approach:
- Cyberbullying and Hate Speech
- Cyber Fraud and Scams
- Reputation Manipulation and Fake Activities
- Propaganda and Radicalization
The main goal of this seminar is to bring together researchers working on all aspects of cybersafety, including security, privacy, human factors, economics, sociology, law, and psychology. We aim to discuss many facets of the problem, both technical and nontechnical, and jointly identify measures to advance the state of the art and identify promising research avenues. Examples of issues to be debated include:
- How do we define cyberbullying and online harassment in a way that captures their inherent ambiguities and subjectiveness?
- How do perpetrators of these activities exploit technological tools to increase their effectiveness? How do cyberbullies and online harassers organize and choose targets?
- What are the different types of cyber fraud activities and how might we cluster different types of scams, based on psychological, sociological, situational and technical variables so as to better design countermeasures?
- What data are ethically and socially acceptable to draw upon in detection and prevention of cyber fraud?
- What variables are important in enabling us to distinguish those who have become single or repeat victims from nonvictims?
- What are the current mitigation schemes adopted by social networks to counter reputation manipulation and their limitations?
- What are the economics and legal mechanisms governing fake activities?
- What can be done to make it economically unviable for fraudsters to engage in reputation manipulation and fake activities?
- How does online radicalization happen?
- Are there specific demographics that are more susceptible to being radicalized? How is online radicalization different from other types of online abuse?
The organizers of this seminar acknowledge research funding from the European Union’s Horizon 2020 Research and Innovation programme under the Marie Skłodowska-Curie Grant agreement No 691025. For more information, please visit http://encase.socialcomputing.eu/
The Dagstuhl Seminar 17372 "Cybersafety in Modern Online Social Networks" was a short two and a half day seminar, which took place September 10th-13th, 2017. Its main goal was to bring together researchers from various research areas related to cyberfraud and cybersafety in online social network, and to inspire them to exchange results, practical requirements, and ethical/legal implications related to user-driven research.
First Day. The seminar started with a short self-introduction of all the participants, then, we had an initial brainstorming session to identify main topics of interest, various aspects involved in them, and a balance in terms of interdisciplinary representation. Specifically, we focused on scams happening in online social network and hate speech, while paying special attention to the protection of minors. The aspects discussed were related to algorithmic, user, understanding/modeling, ethical, and privacy aspects of working in this line of research.
The brainstorming session concluded with the discussion of the following tangible research directions:
- We should work on detection, prevention, and mitigation of hate speech.
- All solutions should be in accordance of regulations.
- We should pay particular attention to false positives, as users can easily lose their trust in the platform.
- We should take into consideration the role of proxies, which act as biases on the data.
- We should focus on counter-terrorism research that aims to distinguish vulnerable population in order to recruit them for propaganda purposes.
We then had four long talks throughout the day. The first speaker, Jeremy Blackburn (University of Alabama at Birmingham, US), described his work on cyberbullying and hate speech that includes studying behavior on video games as well as fringe Web communities like 4chan. The second speaker, Filippo Menczer (Indiana University -- Bloomington, US) ,presented how misinformation is spread on Twitter. Specifically, he presented how false information as well as the respective fact-checking efforts are diffused on the Twitter network. The third speaker, Gianluca Stringhini (University College London, GB) presented his work on cyberfraud and scams, focusing on deceptive techniques employed by malicious users in order to scam benign users on online dating sites. The last speaker of the first day was Awais Rashid (Lancaster University, GB), who described his work related to child sex offenders and how he coordinated with Police bodies in order to undertake research on this topic. Also, he presented the ethical considerations when doing research with sensitive data, like those used for this study.
Second Day. The morning of the second day focused on giving an overview of work done on a variety of topics related to the main topics of the seminar (through short talks from the participants). More specifically, Zinaida Benenson (Universität Erlangen - Nürnberg, DE), described her work on spear fishing, where malicious users aim to deceive users by sending email that contain malicious URLs. Then, Michael Sirivianos (Cyprus University of Technology - Lemesos, CY) presented his work on how to combat friend spam by analyzing the underlying network of social rejections. The next talk was by Alexandra Olteanu (IBM TJ Watson Research Center - Yorktown Heights, US), who discussed some preliminary results on work done on hate speech. Srijan Kumar (Stanford University, US) then showed how sockpuppet accounts are used in social networks to change and manipulate the opinions of other users of the platform. Savvas Zannettou (Cyprus University of Technology - Lemesos, CY) presented his research on how news propagates across multiple Web communities, and how to measure their influence. Then, Manuel Egele (Boston University, US), presented COMPA, which is a system that captures the behavioral profile of the user in order to identity possible account compromises. Huy Kang Kim (Korea University - Seoul, KR) talked about malicious users exploiting video games to make money. Next, Oana Goga (MPI-SWS - Saarbrucken, DE) described how online identities can be strengthen by combining multiple weak identities. The last talk was by Julien Freudiger (Apple Computer Inc. - Cupertino, US), who covered public privacy and safety guidelines used at Apple.
The afternoon was dedicated to two parallel working groups focused on discussions around a particular topic, specifically, one was about future directions on hate speech research, and another about ethical considerations that researchers should keep in mind when working with users or user data.
Third Day. The final day of the seminar had two more parallel working groups, one on research directions related to cyberfraud in online social network, and another on algorithmic biases and possible solutions to avoid it. We then had a discussion summarizing the work and the discussion done in the various working groups and ended up with ideas for future events, collaborations, and follow-ups.
Acknowledgments. The organizers of this workshop acknowledge research funding from the European Union's Horizon 2020 Research and Innovation programme under the Marie Sklodowska-Curie Grant Agreement No 691025. The organizers would like to thank the Schloss Dagstuhl for the professional, productive, and enjoyable atmosphere it provides and for their invaluable support. Finally, we are grateful to and Seda Guerses for taking notes during two of the working groups and to Savvas Zannettou for coordinating the writing of this report and taking notes throughout the seminar.
- Zinaida Benenson (Universität Erlangen-Nürnberg, DE) [dblp]
- Jeremy Blackburn (University of Alabama at Birmingham, US) [dblp]
- Emiliano De Cristofaro (University College London, GB) [dblp]
- Julien Dreux (Facebook - London, GB)
- Manuel Egele (Boston University, US) [dblp]
- Julien Freudiger (Apple Computer Inc. - Cupertino, US) [dblp]
- Oana Goga (MPI-SWS - Saarbrücken, DE) [dblp]
- Seda F. Gürses (KU Leuven, BE) [dblp]
- Huy Kang Kim (Korea University - Seoul, KR) [dblp]
- Christiane Kuhn (TU Dresden, DE) [dblp]
- Srijan Kumar (Stanford University, US) [dblp]
- Ilias Leontiadis (Telefónica Research - Barcelona, ES) [dblp]
- Filippo Menczer (Indiana University - Bloomington, US) [dblp]
- Prateek Mittal (Princeton University, US) [dblp]
- Alexandra Olteanu (IBM TJ Watson Research Center - Yorktown Heights, US) [dblp]
- Awais Rashid (Lancaster University, GB) [dblp]
- Ahmad-Reza Sadeghi (TU Darmstadt, DE) [dblp]
- Stefan Schiffner (ENISA - Athens, GR) [dblp]
- Michael Sirivianos (Cyprus University of Technology - Lemesos, CY) [dblp]
- Gianluca Stringhini (University College London, GB) [dblp]
- Thorsten Strufe (TU Dresden, DE) [dblp]
- Savvas Zannettou (Cyprus University of Technology - Lemesos, CY) [dblp]
- Dagstuhl Seminar 19302: Cybersafety Threats - from Deception to Aggression (2019-07-21 - 2019-07-26) (Details)
- security / cryptology
- society / human-computer interaction
- world wide web / internet
- online social networks
- security and privacy
- legal and ethical issues on the web