2
Bots and Social Life

The Twitter bot @everyword was built to tweet each word from the Oxford English Dictionary. From 2007 to 2014, the automated account shared one word every thirty minutes, in alphabetical order. It eventually shared a total of 109,157 words. The bot developed a following of over 100,000 Twitter users, who regularly liked, retweeted, and commented on its content. This account, carrying out its mundane little task, garnered attention from major news outlets, with writeups in the Guardian (Spencer, 2014), The Washington Post (Dewey, 2014), and The New Yorker (Dubbin, 2013). Its creator, the poet and programmer Allison Parrish, explained that the endeavor was a response to people's perceptions of Twitter content as "inane," filled with "people just posting about their sandwiches or whatever"(Fernandez, 2017). How would a bot sharing each English word challenge, critique, or concretize these assumptions about Twitter?

For many, @everyword was both a pointless and poignant exercise - simultaneously repetitive and captivating. It captured the mundane nature of information transmission online, while also subtly revealing the nuances of language and the online experience. MIT professor and computational poet Nick Montfort summed up the bot as "the everyday bot for the everyman. Everyone thinks everything about it, heading everyway to everywhere" (Parrish, 2015). The genius of @everyword was its elucidation of the procedural underpinnings of computer programming, the web, and language, while also provoking deeply human interaction and reflection. It remains an outstanding example of a social bot (which we define as any bot designed to interact directly with human users.)

Social bots interact with average internet users across a wide variety of websites and platforms. You might encounter them in defined roles, such as the automated digital customer service representatives (chatbots) on your bank's website. Or you might encounter bots in a multitude of less-constrained roles on Twitter, as with @everyword. In a piece defining the "bot," Parrish wrote about what she calls the "PUDG model," which defines bots as "procedural, uncreative, data driven, graffiti" (Parrish, 2016). They are procedural in that their "content is automatically generated without human intervention, using a set of predetermined rules and procedures"; they are uncreative in the sense that they produce "writing that concerns itself with categorizing, remixing, and re-enacting pre-existing textual artifacts" in order to "draw out something new and unexpected in the process"; they are data-driven because they can process a great deal of data, much more than earlier examples of procedural writing; and they are graffiti because they are artistic "interventions in a public space (to the extent that, e.g., Twitter can be defined as a public space)." Social bots are thus a mash-up of the computational and the social, the rote, and the surprising.

Online, bots are part of life. Even bots built for less obviously communicative or creative tasks work to form part of the incongruous fabric of internet society. This chapter therefore looks at bots and social life - mostly social bots, but also bots operating on the periphery of society, underpinning cyberspace and the diverse array of digital ecosystems that most of the world traverse on a daily basis. It is situated in this book's broader argument that bots are an integral part of our connected lives the world over. With social bots, as with bots in their other forms, it's easy to forget they exist - they can be almost ambient, particularly when operating in routine customer service roles. Really, though, they are widespread in a variety of roles across the web. They are growing in population and in sophistication.

The core question at the heart of this chapter is: how are bots social? How do they enable interaction and sociality, both intentionally and unintentionally? How are people and bots connected? Why do people personify bots, attributing personality and human characteristics to them? In the following sections we describe the role of bots in global society, referring back to the typology of bots presented in Chapter 1 (Gorwa & Guilbeault, 2018). Broadly, we talk about bots and the internet, and more specifically about bots and social media. We discuss the social role of bots in the domains of journalism, the arts, and dating. We conclude with a discussion of how bots' social role is evolving as innovations in AI and machine learning open up new avenues of social bot output that enable both routine and unexpected behaviors.

Bots and Global Society

The "thing" that we think of as the internet is actually a system, made up of hardware (physical components like cables, servers, and parts of computers, such as monitors or motherboards); software (information or sets of instructions that give directions to the apparatus); and all of the information and interactions take place online. Social media is a similar complex construct that

consists of (a) the information infrastructure and tools used to produce and distribute content that has individual value but reflects shared values; (b) the content that takes the digital form of personal messages, news, ideas, that becomes cultural products; and (c) the people, organizations, and industries that produce and consume both the tools and the content. (Howard & Parks, 2012)

So, a large portion of both the internet writ large and social media is infrastructural - a combination of the physical and organizational components (the hardware and software) that are necessary for these systems to communicate and create a network between machines, between people, and between combinations of the two. In other words, both the internet and social media are media - means of facilitating information sharing and the generation of meaning. As Howard and Parks point out, messages are "cultural products," and the systems that spread them are rich with both social and cultural significance.

Bots are similarly dynamic, especially when we consider them in terms of Parrish's (2016) PUDG model. Indeed, some scholars have argued that bots should be considered, and studied as a whole new form of media (Woolley, 2017). In their simplest forms, bots are simply automated software programs tasked with doing work online. But that exceedingly broad definition ignores people's deeply social, often personified perception of bots. Theories about human-bot relationships and about the nature of bots are explored in Chapter 6. In this chapter, we focus on the social component of the "lives" of bots. What is it about bots that makes people personify them?

We believe that most simply, people personify bots that are used for front-end (direct) communication with both people and other software programs. Forward-facing bots like @everyword are often given names and provided with short descriptions, which seem to ascribe individuality to the bots. Bots on social media platforms like Twitter or Reddit are often given profile pictures and programmed with particular modes of speaking, further cementing perceptions that they are somehow unique. The bots' automated responsiveness (some even leverage AI) invite people to interact with bots as if they have some degree of sentience. Of course, they do not, but bots are (arguably) partially human, for bots are built by people, and they carry the values and - to some extent - the personalities of their builders (Woolley et al., 2018). But, while bots act and interact as extensions of their creators' will and desires, Parrish (2016) notes that bots often do unexpected things that their creators did not intend. This is because they operate online in ecosystems populated by all sorts of other forces and individuals - outside influences that can produce unanticipated behaviors. As Gehl and Bakardjieva (2016) point out, the actions of socially oriented bots can have all sorts of benefits - and they can cause corresponding losses - for users online.

Because bots have all these characteristics - because they are imbued with social meaning, because they are humanized, and because the tasks they undertake are immersed in systems of communication and culture - bots can be seen as enacting social force. What does this mean? First, it means that bots are influential. As we will discuss throughout this book (and at length in the chapter on bots and politics), bots are often used, both directly and indirectly, to persuade computational systems and the people that use them to do certain things. Particularly in relation to politics, this seems somehow Machiavellian or nefarious. But most bot influence is in fact routine or programmatic: most bots aren't explicitly built to manipulate public opinion but to transmit information, but sometimes the act of transmitting information introduces change. Of course, just because there are not always nefarious intentions behind programmatic bots or apparently routine algorithms does not mean that the effects they produce are necessarily positive, or even neutral: As Noble (2018) points out, minority communities often bear the brunt of these oppressive effects.

Second, bots exert social force because they connect to people. Humans perceive bots as part of...