bots

Over the past decade, expanding access to Big Data has produced a number of innovations in electronic literature and digital culture more broadly, ranging from Twitter bots, media art and generative poetry utilizing social data to vernacular creative writing, journalism and fictocriticism on platforms such as Tumblr and BuzzFeed. These divergent modes of expression all rely on the ability to find and sort high-volume, real-time, multimodal digital data – for example tweets, Instagram photos, animated GIFs, YouTube videos, SoundCloud audio tracks and more – and recombine them in novel works of bricolage. Yet despite the increasing prominence of these writing practices, they have received scant scholarly attention.

In this paper, I propose that we consider these works as a discrete class that employ a novel and distinctive orientation to literary craft: namely, the central writerly act inheres not in the crafting of sentences but in interfacing with data structures via search string manipulation. This process typically has three steps: first, a search string is constructed with primarily Boolean operators; next, the results are sorted via manual browsing or algorithmic filtering; finally, the disparate content is assembled together with traditionally crafted text or paratext into a coherent whole. I demonstrate the technique with two examples: the data-driven e-lit installation “Death of an Alchemist” by myself and Dr Andrew Burrell, appearing at ISEA2015, and the “Buzzademia” digital humanities initiative led by Mark Marino, which I have been closely involved in as a writer for BuzzFeed.

With its emphasis on repurposing online content, this emergent digital writing technique clearly must be understood as belonging to the broader ecosystem of remix culture; it also has obvious links to the conceptual poetry movement. However, the emphasis on optimizing data search sets these works apart from those related tendencies. Invoking Sigmund Freud’s analogy of the “mystic writing pad”, which has previously been compared with hypertext, I suggest that we understand this new poetics through a related metaphor: scratch art paper, a children’s toy that allows the user to trace an original figure that is wholly constituted by another, previously created drawing. Literary originality is, increasingly, expressed through the deployment of virtuosic search terms aimed at finding the creative work of others.

This nascent form of poetics is, I argue, a defining literary technique of the age of Big Data. Indeed, for several reasons, we may consider such writing as being not literary but post-literary. It often resists categorization under the rubric of “literature”, proudly associating with lowbrow and vernacular forms of communication. Such writing also eschews traditional models of literary authorship in favour of a liminal form of human-machinic agency. Finally, it is often pervasively multimodal, de-emphasizing the written word in favor of image, video and other non-verbal data.

The post-literary turn, if we accept that is what these forms of writing represent, offers some exciting new modes of creative expression. On the other hand, it may also be considered symptomatic of what Peter Sloterdijk has called the waning power of language – and the growing tyranny of images and data – under late capitalism. Reconceiving search strings as literature thus presents a tangled knot of opportunities and problems.

(source: ELO 2015 Conference Catalog)

With half a century’s worth of profound social and technological change, the 1960s protest movement is far removed from today’s world. Networks, databases, video games, social media, and the rise of algorithmic culture and the sharing economy have irrevocably altered our landscape. What, in this world, is the 21st century equivalent of that key feature of the sixties protest movement: the protest song? This paper argues that one possible answer is the protest bot, a computer program that algorithmically generates social and political critiques on social media.

Using Habermas’s imperfect account of the public sphere as my starting point, I suggest that five characteristics define protest bots—or bots of conviction, as I also call them. Bots of conviction are topical, data-based, cumulative, oppositional, and uncanny. After explaining these five characteristics, I explore several well-known and lesser-known bots on Twitter, showing how they are or are not protest bots. Throughout this paper I adopt a critical code studies approach, diving into the procedural DNA of several bots of conviction of my own creation.

This paper situates bots of conviction within a larger bot ecology, which includes spambots, chatbots, generative poetry bots, art bots, and absurdist avant-garde bots. Unlike these other forms of computer-generated social media, protest bots are a promising form of tactical media, a kind of media activism that destabilizes dominant narratives, perspectives, and events. This micropolitical activism, I argue, should complement the creative dimensions of digital art, poetry, and literature, serving as one of “the ends” of electronic literature.

(source: ELO 2015 conference catalog)

Language is the hidden scaffolding of networks, applications, and web sites. It is minified and monetized in ways that are often occluded from the everyday user’s experience. From their point of view, the interaction is innocuous – language is used for labels and explanations. A few words are typed into an empty field and thousands of related results appear instantly. A simple search, an email to a friend, a unique phrase – all easily logged, monetized, and indexed. This is the world of invisible participation.

Our panel is interested in language on the Internet, how it is created, by whom, where it exists, and how it is used. Three examples: Google reads our emails, garners information from our personal messages and uses that profiling strategy to select “relevant” ads. It then displays those ads on the screen next to the very emails from which the information was initially taken. Facebook and other social media platforms use similar methods of securing and storing data — data that is paradoxically private and public, and all personal. Further, crowd-sourced encyclopedias like Wikipedia are shaping the way we read, learn, and think. Language is what links all of these sites together. All of the sites’ underlying organization and structures have been built to follow the logic we ourselves employ in using language. “Robots” read content, algorithms interpret it and databases memorize it. The impact of this process is no longer confined to the Internet, but has reached beyond it into our everyday lives.