Socialbots and Their Friends

Digital Media and the Automation of Sociality

Robert W. Gehl, Maria Bakardjieva, Robert W. Gehl, Maria Bakardjieva

• Publisher: Routledge
• Year: 2016
• ISBN:
  9781317267386
• Edition: 1
• Topic: Computer Science
• Subtopic: Human-Computer Interaction
• Index: Computer Science

Many users of the Internet are aware of bots: automated programs that work behind the scenes to come up with search suggestions, check the weather, filter emails, or clean up Wikipedia entries. More recently, a new software robot has been making its presence felt in social media sites such as Facebook and Twitter – the socialbot. However, unlike other bots, socialbots are built to appear human. While a weatherbot will tell you if it's sunny and a spambot will incessantly peddle Viagra, socialbots will ask you questions, have conversations, like your posts, retweet you, and become your friend. All the while, if they're well-programmed, you won't know that you're tweeting and friending with a robot.

Who benefits from the use of software robots? Who loses? Does a bot deserve rights? Who pulls the strings of these bots? Who has the right to know what about them? What does it mean to be intelligent? What does it mean to be a friend? Socialbots and Their Friends: Digital Media and the Automation of Sociality is one of the first academic collections to critically consider the socialbot and tackle these pressing questions.

1
Socialbots and their Friends

Robert W. Gehl and Maria Bakardjieva

The goal of this book is to draw attention to the arrival of a new socio-technical entity—the socialbot. This arrival has not been bombastic, but rather quiet for now. The first sightings of the socialbot across the digital media field date from only a few years ago. It remains an evasive creature and an object of fantasies and rumours. Nobody knows for sure where it will turn up next, how it will figure in established social practices and even less in what way it may change them. The meaning of the socialbot is being invented in laboratories, appraised in the marketplace and construed in the mass media as we speak.

This is exactly the reason why we believe social and cultural researchers should take the time to speak about it. Our prediction is that the socialbot will soon become a customary inhabitant of our shared media environment. As scholars, we are willing to be active participants in the cultural construction of the social-bot and intend to mobilize our experience with genealogy, ethnography, critical theory, and technology and media studies to understand its nature, prospects and potential social significance.

We start with a crisp technical definition offered by some of the engineers whose pioneering research raised attention to this phenomenon and put it on the map:

A socialbot is an automation software that controls an account on a particular OSN [Online Social Network], and has the ability to perform basic activities such as posting a message and sending a connection request. What makes a socialbot different from self-declared bots (e.g., Twitter bots that post up-to-date weather forecasts) and spambots is that it is designed to be stealthy, that is, it is able to pass itself off as a human being.

(Boshmaf et al., 2011, p. 93)

What aspects of this definition taken from a computer science journal could possibly spark the curiosity of the students of media, society and culture? At first glance, as a type of automation software, the socialbot does not herald anything new and exciting. The fact that it controls an account on a particular online social network brings it closer to a field of investigation popular among social researchers, but this field is already infused with rather trite automation software (bots) tasked with all kinds of repetitive activities and chores.

However, the fact that the socialbot is by definition ‘designed to be stealthy’, and is typically willing and ‘able to pass itself off as a human being’ signals a significant change in the game. From the perspective of the social and cultural analyst, this fact is remarkable because it indicates that the socialbot is designed not simply to perform undesirable labour (like spambots) and not only to try to emulate human conversational intelligence (like chatbots). Rather, it is intended to present a Self, to pose as an alter-ego, as a subject with personal biography, stock of knowledge, emotions and body, as a social counterpart, as someone like me, the user, with whom I could build a social relationship. The context of online social networking sites represents both a defining characteristic and a condition sine qua non for the existence of socialbots. Without the conventional interface and functionality of social networking platforms, the fabrication of a believable resemblance of a human Self and the interactions with human users on the part of the socialbot would have been unthinkable. Thus, the socialbot, a kind of automation software equipped with the ability to act and pass itself off as a human being and a social Self on social networking platforms, opens a new frontier of human experience, that of robo-sociality. Our book sets itself the goal of providing some early glimpses on the developments and challenges emerging on this frontier.

In an extended version, our working definition of socialbots for the purposes of this book can be formulated as follows:

Socialbots are software processes that are programmed to appear to be human-generated within the context of social networking sites (SNSs) such as Facebook and Twitter. They achieve their ‘humanness’ by either mimicking other SNS users or through artificial intelligence that simulates human users of social networking sites. They share pictures, post status updates and Tweets, enter into conversations with other SNS users, and make and accept friend and follower requests. Importantly, they are designed to appear human to both SNS users as well as the SNS platform itself. Their goals are various, but often include shaping the online interactions and social networking practices of users.

Thus, we are specifically referring to automation software that operates within social networking sites and is purposefully designed to appear to be human. As with all technological artefacts, the socialbot does not burst out of the blue, but represents a product of a long and winding history of intentions, experiments and inventions (see Williams, 1974). It descends from a long lineage of conceptual and technical precursors.

However, we will resist the temptation to construe the socialbot’s relationship to these earlier contraptions with similar features, purposes or inclinations as a family tree or an evolutionary chain. Instead, drawing on the language of the social networking sites where socialbots characteristically dwell, we refer to their conceptual precursors as ‘friends’. To pick up a term made trendy by Mark Zuckerberg, we are thinking of a ‘social graph’ of algorithmic processes that bear the marks of mimetic kinship; a map of connections between software agents exhibiting commonality, congeniality and co-orientation.

Among the ‘friends’ who clearly connect to socialbots in one way or another we find a whole range of software agents, including chatbots such as ELIZA, Cleverbot and MrMind (see the chapter by Weil in this book); intelligent agents such as Cortana and Siri (see Guzman’s chapter); utility programs that clean up Wikipedia and Reddit (see Massanari’s chapter); recommendation engines; Web spiders and search engines; and spambots. Some of these agents have, like the socialbot, been able to pass as human within their specific sociotechnical contexts: chatbots in Multi-User Dungeons (MUDs), bots in Internet Relay Chat (see Latzko-Toth’s chapter), or automatic avatars in Second Life (see Muhle’s chapter). One example of a bot passing as human is Julia, a MUD bot that, according to Sherry Turkle, ‘is able to fool some of the [other MUD users] some of the time into thinking she is a human player’ (Turkle, 1995, p. 88). At a higher level of complexity, David Burden has attempted to create automated ‘robotars’ in Second Life that can express emotions and react to stimuli within that virtual world (Burden, 2009; Slater & Burden, 2009).

So what gives socialbots a special place in this social graph? We suggest that what is unique about them is how they perform socially within the specific contours of social networking sites. They have profiles, they like, friend, follow and tweet, and they carry on conversations. They are, overall, capable of operating within one of the most popular communication practices of our era. This book is intended to be a first pass at mapping this algorithmic social graph; our desire is to articulate socialbots into their networks of software-based—and human—friends, and to consider their current and future role within the socio-technical assemblages that are contemporary SNSs.

This social graph of automated sociality is a complex one. Although there appears to be a clear boundary between a computational process that can intentionally simulate humanness and an intelligent agent (say, Siri) that does not intend such deceit, as the chapters of this book show, the boundaries of both ‘botness’ (Bucher, 2014) and ‘humanness’ are blurry. Indeed, as Robert Gehl (2014), Alexandra Samuel (2015) and Maria Bakardjieva (see her chapter) have pointed out, SNSs privilege the rationalization and automation of human activity, making even human-to-human communication quite bot-like, or ‘cyborg’ to use a term from philosopher Donna Haraway as well as computer scientists Chu et al. (2010). Socialbots may present a new and somewhat extreme case of the old problem of robot/human confusion, but this case emerges against the background of long-standing and evolving challenges to anthropocentric ontologies as the stories of the earlier bots detailed in this book demonstrate. In sum, both socialbots and their friends challenge us to think about what it means to be human and to be social in a time of intelligent machines.

Socialbots Timeline

This timeline is not meant to be an exhaustive historical account of all social robotics—that is, robots of all forms (software and mechanical) that are meant to socially relate to humans. For that history, see Chemers (2013); for a (an earlier, but still valuable) survey of contemporary social robotics, see Fong et al. (2003). Instead, our goal is to capture some of the notable events in the young lives of the representatives of the particular species of automation software that we define as socialbots. They can be seen as instances marking the ‘coming out’ of socialbots on the scene of public awareness and researchers’ attention.

2008

Project Realboy

Run by Zack Coburn and Greg Marra, Project Realboy was an experiment in making Twitter bots that would clone other Twitter users’ tweets, follow other Twitter users and get at least 25% of those followed users to follow back (Coburn & Marra, 2008). As Coburn and Marra note, this is a process of ‘imitating’, rather than ‘impersonating’, other Twitter users. ‘Imitation’ is a key method of socialbots, which tend to reflect our sociality back at us, rather than simply copying a particular human’s social media use.

2009

Twitter Gets Automated

Cheng and Evans (2009) sampled 11.5 million Twitter accounts between January and May 2009 and discovered that 24% of all tweets generated in that period were coming from automated accounts posting over 150 tweets a day. Their study includes data on the top 5% of Twitter accounts by number of tweets, many of which are machines, and also on social media marketers, identified by keywords contained in their profile descriptions. The top 5% accounted for 75% of all activity; 35% of the social media marketers tweeted at least once a day, compared with 15% of all accounts.

2011

Web Ecology Challenge

Based in Boston and run by Tim Hwang, the Web Ecology Project hosted a competition for socialbots to invade Twitter for the chance to win ‘$500 hoo-man dollars’ (‘Competition’, 2011). These socialbots had to pass as human and scored points for getting the most interactions from other (ostensibly human) Twitter users. The socialbot that won this contest was James M. Titus, ‘a Kiwi, living in Christchurch who was obsessed with his pet cat, Benson’ (Aerofade, 2011).

Boshmaf et al. Described their Facebook Socialbot Experiment

Meanwhile, on Facebook, computer security researchers Boshmaf et al. (2011) activated a predacious socialbot network. Using this network, they were able to sign up friends and—working around Facebook’s privacy settings–gather 250GB of private data.

2012

Institute for the Future/Pacsocial Bot Challenge

On the West Coast, Tim Hwang and his colleagues at the Pacific Social Architecting company worked with the Institute For The Future to hold another social-bot competition (Weidinger, 2012). The winner was @Ecartomony, ‘a business school graduate bot with a “strong interest in post-modern art theory”’ (McMillan, 2012), who tweeted provocative, if somewhat jargon-laden, art theory ideas.

Socialbot Experiment, UW-Bothell

North of the Institute For The Future, another Twitter socialbot competition was conducted at the University of Washington-Bothell (Allaouchiche et al., 2012).

2013

Carina Santos

Researchers discovered that Carina Santos, a highly influential Twitter user and reporter, was actually a socialbot (Urbina, 2013).

2015

DARPA SMISC Bot Detection Challenge

DARPA’s Social Media in Strategic Communications held a month-long competition for researchers to build systems to more accurately identify social-bots in Twitter (Subrahmanian et al., 2016). The justification was that such bots can have an undue political influence in SNSs.

Previous Scholarship

As should be clear from our timeline, socialbots are a relatively new invention. They are still in search of their true calling and somewhat sporadically show up in various contexts. They have a dual existence as a more or less furtive undertaking by commercial, political and other operators in various fields of practice, on the one hand, and as an object of premonition and experimentation on the part of scientists and programmers, on the other. In the latter capacity, they have the potential to become a prime example of self-fulfilling prophecy. Existing socialbot scholarship is located predominantly in the area of computer science. Two main threads can be discerned in it: information and network security research; and research on the capacities of socialbots to shape online networks and communication.

Computer Security Research

Information security researchers were the first to name and explore socialbots in the sense of our definition (Boshmaf et al., 2011). Situated in the field of computer security, the language of these articles is harsh: they are stories of attacks and defence, susceptibility and immunity, exploits, penetration and invasion, with socialbots seeking to defeat automated detection systems and mislead users by infiltrating social networks.

Boshmaf et al. built socialbots to test the capabilities of the Facebook Immune System (FIS), a process designed to purge Facebook ...

Table of contents

• Cover
• Title
• Copyright
• CONTENTS
• List of Figures and Table
• Preface
• Author Bios
• 1 Socialbots and Their Friends
• PART I Friends
• PART II Socialbots
• Index