Contributor: Rachael Graham Lussos
Affiliation: George Mason University
Email: rgraham4 at gmu.edu
Released: 19 December 2017
Published: Spring 2018 (Issue 22.2)
This article describes four benefits and two challenges of teaching Twitter bots as a form of digital writing, drawing on the literature of critical making and activist composition assignments. An observation and a survey of students completing Twitter bots for a graduate-level seminar provides further insights into the benefits and challenges of teaching Twitter bots as a form of digital writing.
Twitter bots are a non-traditional medium for academic writing, even for academic digital writing. Twitter bots are Twitter accounts that perform programmed functions without human intervention, for practical, playful, or serious purposes. A bot that automatically tweets weather alert updates is both practical and serious, but a bot like Darius Kazemi’s @TwoHeadlines, which continuously pulls two headlines from Google News and combines parts of each to create a new headline, is more playful (see Figure 1). The brevity of Twitter bot messages is unusual in academic writing: the textual element of a Twitter bot’s message, if it even has text, comprises 280 characters or fewer. (Prior to November 2017, the limit was 140 characters.) As a digital writing medium, the delivery of Twitter bots' messages makes Twitter bots that much more unique: the endless and automated composition and production is unlike the delivery of any traditional academic assignment. That is, to compose a Twitter bot, a writer must first plan the function that the bot will perform; second, write a program (either from scratch or using an open-source program, as I describe further below) that enables the bot to perform the envisioned function; third, test and tweak the program until the bot performs the function to the writer's satisfaction; and finally, launch the program and allow the bot to perform its programmed function with no or limited interference from the human writer.¹ Therefore, to compose a Twitter bot, students must first conduct rhetorical analyses of purpose, audience, and delivery before they can begin “writing” (i.e., programming) the final product.
Transcript: Two screenshots from Twitter. The first screenshot shows the Twitter profile page for the account "DC Weather Alerts," with the Twitter handle @dcWeatherAlerts. The profile image is a drawing of the dome of the U.S. Capitol Building, with clear and stormy weather on either side of the dome. The bio reads, "It's like @capitalweather, but only giving you the most important of updates. This is an automated feed." The account location is Washington, DC and start date is February 2009. The account summary is 2,683 tweets, 5 accounts followed, and 10.5K followers. One tweet is shown, linked to The Washington Post and an image of the Lincoln Memorial. The tweet reads, "Severe Thunderstorm Warning for Fauquier and Loudoun County until 6:00pm." The second screenshot shows three tweets, by a bot named "Two Headlines," with the Twitter handle @TwoHeadlines and a profile picture of a lithograph of a printing press. From top to bottom, the tweets read, "PM Tom Brady justifies demonetisation amid criticism," "Xbox One had plenty of big hits from 1879 to 1950," and "President Trump is now holding the Amazon.com, Inc. hostage."
While others have examined how and why writing instructors incorporate social media technologies in first-year composition (Buck, 2015; Miller et al., 2015) or technical communication (Daer & Potts, 2008), the focus of this webtext is intentionally limited to writing with a social media technology as it relates to critical making and, thereby, as a medium for activism. Furthermore, expanding on this previous literature, which examined writing by human users, this webtext focuses on writing that is programmed by humans to be arranged and delivered by automated programs. This focus on programming as writing addresses a call for teaching “computational literacy,” in which instructors help make aware “the writing practices that undergird our complex, contemporary composition environments” (Vee, 2013, p. 58). By writing Twitter bots, students can experience how the hidden writing of social media technologies—the automated programs that enable (or in some cases, disable) use of those technologies—involves a rhetorical analysis.
In this webtext, I first describe four benefits of using Twitter bots in teaching digital writing before discussing two challenges. I then describe student reception of a Twitter Bot assignment in a graduate course I observed. I conclude with recommendations for teachers and researchers who plan to teach or study Twitter Bots in digital writing contexts.
Twitter bots share the benefits that generally apply to all critical making assignments. Critical making assignments are assignments that invite students to construct an argument using a non-traditional medium, which can be anything from needlepoint to decorated cakes. As defined by Matt Ratto and Megan Boler (2014) in the introduction to their edited collection DIY Citizenship, critical making is “an activity that provides both the possibility to intervene substantively in systems of authority and power and that offers an important site for reflecting on how such power is constituted….Critical making also signals the integration/simultaneity of processes and practices, the act of making ‘things’” (emphasis theirs, p. 1). This definition of critical making highlights two benefits of critical making assignments: first, an opportunity to analyze and critique (and possibly intervene in) dominant systems and structures of power (which I discuss further under Benefit Two), and second, an explicit focus on process and practice.
This explicit focus on process and practice is key to critical making assignments. Working with an unfamiliar medium, such as needlepoint, cakes, or bots, forces students to question much more deliberately the processes of how to compose with that medium. Then, the awareness of those processes forces the necessary rhetorical analysis to the forefront—analysis that can easily be overlooked when one works with a more familiar medium, such as writing for an essay. In Toward a Composition Made Whole, Jody Shipka (2011) described how a close examination of the composition process—which happens when one composes with a new medium—enables one “to make the complex and highly distributed processes involved with the production, reception, circulation, and valuation of texts more visible” (p. 38). For example, with Twitter bots, the writer of the bot must not only analyze the audience of the bot's messages but also analyze how the intended audience will access the bot's messages, and this analysis must be complete before the bot tweets even a single message. The writer of the bot might decide that the bot will engage with its intended audience by replying to tweets containing certain keywords, or the writer might simply have the bot follow the Twitter accounts of its intended audience. Either of these methods for engaging with an audience necessitates a completely different approach to designing, programming, and launching the Twitter bot. Therefore, students writing Twitter bots must explicitly analyze the processes of "production, reception, circulation, and valuation" before they begin the actual "writing" of the bot.
Twitter bots resolve a common limitation of critical making assignments. Because of the nature of critical making projects, many of these projects are limited to an audience of one’s classmates and teacher, whereas some projects might access a larger audience as a campus or art exhibit installation. However, there are many arguments to be made that are intended for an audience that is not physically accessible from a college campus or art exhibit. Networked mediums such as Twitter bots circumvent this issue by granting access to a large and diverse audience on a social media platform.
Digital studies scholars and expert bot-makers Mark Sample and Zach Whalen recognize the activist potential of Twitter bots, identifying a genre of Twitter bots as "protest bots." Sample described protest bots as "the 21st century equivalent of a protest song," a modern version of the 19th century "journalism of conviction" explored by Jürgen Habermas in his theories on public discourse. Citing Sample's article, Whalen explained how the space of Twitter affords the opportunity for "making something visible that’s not supposed to be visible" (2014). Other examples of the political relevance of Twitter bots, such as bots that influence elections (Alfonso, 2012; Sharkov, 2017), demonstrate that creating an activist bot or protest bot is not only feasible but, potentially, adds a necessary voice to this influential public space.²
Because Twitter bots function in an easily accessed public space, they are an ideal critical making assignment for communicating messages not only to a larger audience but more importantly, to their intended audience, which is especially important for activist messages. Being able to access one’s intended audience makes it that much easier for students to imagine writing to a “real” audience in the first place, one of the cited benefits of activist composition (Ervin, 1997). Just as the non-traditional delivery of Twitter bots helps students deliberately analyze the rhetorical implications of delivery, being able to reach the intended audience of an activist message will help students deliberately analyze audience.
The risks and challenges of protesting in physical spaces has been documented by instructors who assign activist composition assignments (Welch, 2005; Feigenbaum 2012), but protests in online spaces can also elicit threats to one’s mental, emotional, and physical well-being (Collins, 2016; O’Brien, 2016; Wachs, 2016). Being able to preserve anonymity during a protest is especially helpful for online activism.
In addition to explicitly activist Twitter bots, Joshua Daniel-Wariya (2016) cautioned that “any concept of play for rhetoric and composition should acknowledge the risk involved when composing, because the act of composing nearly always entails some form of risk” (p. 34). Whether students compose bots that protest or merely critique, the anonymity afforded by Twitter bots helps protect students from a potentially volatile public space.
The experience of designing a bot requires that the designer develop a foundational understanding of how programming works, an increasingly desired skill. The benefits of understanding programming include enhanced creative thinking and logic skills as well as the pragmatic knowledge for automating technological functions (Hooda, 2017; McFarland, 2014). In addition to industry pleas for a workforce who understands programming, in "Understanding Computer Programming as a Literacy," Annette Vee (2013) addressed the importance of developing this understanding specifically for those in writing studies: "programming is leaving the exclusive domain of computer science and becoming more central to professions like journalism, biology, design—and, through the digital humanities, even the study of literature and history" (p. 46). As Vee explains, writing studies scholars do not have to master a programming language to benefit from developing an understanding of how computer programs are written. In the very least, writing studies scholars and students can benefit from understanding how computer programs influence and are influenced by humans and social contexts.
Designing and launching a Twitter bot with an open-source program can provide the most basic introduction to how programming works, without even having to learn a programming language. In the least, students get a glimpse of the type of logic skills they need in order to compose a series of commands to achieve a certain function. Plus, knowing how these logic skills must precede programming is a nice analogy for understanding how rhetorical analysis must precede writing.
Assessment is a common challenge for critical making assignments, including Twitter bots. In “Assessing Scholarly Multimedia: A Rhetorical Genre Studies Approach," Cheryl Ball (2012) offered a helpful rubric for assessing multimedia assignments, which might apply well to the assessment of Twitter bot composition. However, if you assign the composition of a Twitter bot that is explicitly activist in nature, consider using Mark Sample’s (2014) taxonomy for protest bots as a heuristic. For example, you can have students write reflections on how their bot did or did not meet each of Sample’s five criteria for a protest bot: topical, data-based, cumulative, oppositional, and uncanny. You can then assess the extent to which the students demonstrated their comprehension and analysis of Sample’s text in relation to their own composition process.
There is not much of a middle ground between simple formulas you can program in an hour and more complex and unique bot designs that require multiple hours of coding, especially if the composer is unfamiliar with programming languages. Unless you plan to push your students to learn some actual programming, I recommend that you become familiar with the limitations of open-source programming, so you can set some reasonable expectations for the types of arguments your students can compose.
To test these assumptions about the benefits and challenges of assigning the composition of Twitter bots, I observed one class in a graduate-level seminar titled "Writing and Rhetoric in the Age of Algorithms" and surveyed the students. The goal of the course was to learn how algorithms function rhetorically and how to engage in algorithmic reading and writing. Initially, students were required to design and launch a bot by the end of the course, but this assignment eventually became optional.
During the class I observed, the instructor led a discussion about Sample's article on protest bots, which the instructor had assigned previously as required reading. After discussion, students worked in groups to choose a topic of protest and envision a protest bot design for that topic, loosely imitating the goals of Sample's @NRATally bot. In the remaining few minutes of class, students worked individually to brainstorm different types of bots (protest bots or otherwise) they might like to build, considering topics, audiences, methods for getting data, and structure.
After the course ended, the instructor sent an anonymous online survey I created to the course’s 14 students. Six students completed the survey.
My findings from the observation and the survey enhanced my understanding of the benefits and challenges of teaching Twitter bot composition in two ways, both of which primarily concern the activist potential of Twitter bot assignments.
As I describe under Benefit Two, one of the benefits of Twitter bots as critical making assignments is their access to audiences outside the classroom, including intended audiences of protests. However, despite reviewing Mark Sample’s taxonomy for protest bots, students in the graduate-level seminar struggled with and resisted the idea of Twitter bots as agents of activism. Criticisms included describing the activism performed by Twitter bots as akin to “slacktivism.”
My recommendation to instructors who want to use Twitter bots as activist critical making assignments is to have a deliberate classroom discussion of the different types and goals of activism, including slacktivism.³ Discuss whether it is enough for an activist bot to raise awareness or if a bot must also intervene, and review bots that have performed either task. For example, Sample’s @NRA_Tally bot raises awareness of the NRA’s rhetorical strategy for responding to gun-related tragedies but does not engage with any opponents of gun control, whereas the @she_not_he bot raises awareness of transphobic microagressions and intervenes, by automatically replying to tweets that misgender Caitlyn Jenner (see Figure 2) (Dewey, 2015). That said, composing Twitter bots for reasons other than activism has other benefits, as I described previously and will discuss further.
Transcript: Two screenshots of tweets. The first screenshot shows three tweets by a bot named "NRA Tally," with the Twitter handle @NRA_Tally and a profile picture of a shooting target shaped like a human. From top to bottom, the tweets read, "9 tourists shot dead in San Ysidro with a 9mm SIG Sauer P226. The NRA blames mental health system," "13 hospital personnel fatally shot in Royal Oak, Michigan with a 9mm Ruger SR9 semiautomatic. The NRA blames gun control laws," and "16 teachers and students shot dead in Connecticut with a .38-caliber Colt revolver. The NRA says blood is the cost of freedom." The second screenshot shows three tweets by a bot named "She not he," with the Twitter handle @she_not_he and a profile picture of Caitlyn Jenner's face from the cover of Vanity Fair. From top to bottom, the tweets read, ".@tonyburkett ~ROBOT CLANKING NOISE~ It’s she, not he," ".@harrypooper6 Beep beep! It’s she, not he," and ".@cordyalexander Beep beep! It’s she, not he."
Students’ issues with the concept of activist Twitter bots seemed to be tied to the limitations of open-source programming, discussed under Challenge Two. During the classroom observation, students worked in groups to brainstorm ideas for activist Twitter bots. Topics included immigration and deportation, climate change, teacher salaries, standardized testing, food insecurity, and big game hunting. Although the students had little trouble identifying topics for an activist Twitter bot and even sketching out what such a bot might do, many of the students expressed frustrations with the limitations of an open-source program. (Note: The instructor previously assigned the class to review Zach Whalen’s instructions for developing a bot using Google Spreadsheets. See Figure 3 for images of the process using Whalen's instructions, to create the bot @2care4hair, a critique of beauty standards in hair culture.) Meanwhile, most students also expressed their discomfort in engaging with any kind of programming, open-source or otherwise. Working without an open-source program seemed impossible for most of the students, and yet, most of their ideas for activist Twitter bots required something more specialized than they could execute with an open-source program.
Transcript: Series of six screenshots. Top left image shows a section of the homepage of Zach Whalen's instructions for making a Twitter bot using Google Spreadsheets, Version 4.0. The top right image shows the top of a page in Google Spreadsheets and has the headline "SSBot" and the subtitle "A Twitter Bot Engine in Google Spreadsheets." The middle left image shows the middle of a page in Google Spreadsheets with small black text and highlighted green cells for entering text. The middle right image shows a Google Spreadsheets page with the heading "Preview Output" and small black text in a series of rows. The bottom left image shows a Google Spreadsheets page with the heading "Select One from Each Column" and small black text in one long blue shaded row above four small columns partially filled with small black text. The bottom right image shows three tweets by a bot named "Hair Care," with the Twitter handle @2care4hair and a profile picture of the top of a person's head. From top to bottom, the tweets read, "gray hair DO care," "armpit hair don't care," and "dark magenta hair DO care."
The instructor made “activism” an optional feature of the final Twitter bot assignment, and whether it was for issues of programming limitations or negative perceptions of slacktivism, the six survey responses showed that none of the respondents agreed with the statement, “I identify my Twitter bot as activist.” That said, none of the survey respondents disagreed with the statement, “I composed a Twitter bot that focuses on a topic that I care about” (see Figure 4).
Transcript: A screenshot of three tweets, by a bot named "Not Anish Kapoor," with the Twitter handle @notanishkapoor and a profile picture of a jar of pink dye and the words "I AM NOT ANISH KAPOOR" in plastic. From top to bottom, the tweets read, "I own aqua. Painters may not use. #IamnotAnishKapoor," "I own turquoise. Painters may not use. #IamnotAnishKapoor," and "I own metallic red. Advisors may not use. #IamnotAnishKapoor."
Although this classroom experience did not demonstrate “activist potential” as a benefit of Twitter bot composition assignments, the survey responses showed largely positive reviews of the assignment overall. One-third of the respondents reported understanding programming logic better after the assignment, and the majority of respondents agreed that making a Twitter bot was fun and was not a waste of time.
Because Twitter bots are critical making assignments, they are implicitly connected with “the possibility to intervene substantively” as activist agents (Ratto & Boler, 2014), but as the classroom observation and survey showed, more work is needed to realize this activist potential. Specifically, conducting this type of observation and survey with undergraduate composition students, such as in a first-year composition course, will likely yield different results, for better or worse, than the graduate-level summer course observed for this study. Finally, I recommend that teachers using Twitter bots as digital writing assignments regularly update their knowledge of the available open-source programs for Twitter bots, given that new versions and updates are released every few months, often adding more functionality and mitigating previous challenges.
My utmost thanks goes to Steve Holmes for his mentorship during my investigations of Twitter bot composition and for allowing me to observe and survey his class, to whom I am also extremely grateful for their patience with me and generosity with their time. I also thank the reviewers and the PraxisWiki editors for their thorough and kind feedback, which helped improve the quality and clarity of the article.
¹ The rules governing the extent to which a human writer can intervene with a bot's performance are subjective. For example, a writer might eventually edit the bot's programming to change the frequency with which the bot tweets. Exhibiting a greater level of human interference, a writer might manually follow certain Twitter accounts on behalf of the bot's account.
² For an extended discussion on how, why, and how well protest bots and activist bots function, see "Cultivating Metanoia in Twitter Publics: Analyzing and Producing Bots of Protest in the #GamerGate Controversy" by Steve Holmes and me, in Computers and Composition (2018).
³ For a potential reading assignment on the ethics of slacktivism, see Jared Colton and Steve Holmes’s chapter on “Generosity in Social Media Technology” in Rhetoric, Technology, and the Virtues (2017).
Alfonso, Fernando III. (2012, May 21). Twitter bots silence critics of Mexico’s leading presidential candidate. The Daily Dot. Retrieved November 27, 2017 from https://www.dailydot.com/news/pena-nieto-twitter-bots-mexico-election/
Ball, Cheryl E. (2012). Assessing scholarly multimedia: A rhetorical genre studies approach. Technical Communication Quarterly, 21(1), 61‐77.
Buck, Elisabeth H. (2015). Facebook, Instagram, and Twitter—oh my!: Assessing the efficacy of the rhetorical composing situation with FYC students as advanced social media practitioners. Kairos, 9(3). Retrieved October 31, 2017 from http://technorhetoric.net/19.3/praxis/buck/classroomactivities.html
Collins, Ben. (2016, June 20). Twitter trolls are reporting pro-LGBT Muslim women to their governments, where punishment can mean death. The Daily Beast. Retrieved May 28, 2017, from http://www.thedailybeast.com/articles/2016/06/20/the-twitter-trolls-terrorizing-muslim-girls
Dewey, Caitlin. (2015, June 2). I created the Caitlyn Jenner bot @she_not_he. This is what I learned. The Washington Post. Retrieved August 9, 2017, from https://www.washingtonpost.com/news/the-intersect/wp/2015/06/02/i-created-the-caitlyn-jenner-bot-she_not_he-this-is-what-i-learned/?utm_term=.1fac116b31a4
Daer, Alice R., & Potts, Liza. (2008). Teaching and learning with social media. Associate Editor, 21. Retrived October 31, 2017 from http://www.cptsc.org/pp/vol6-2/vol6-2.pdf#page=25
Daniel-Wariya, Joshua. (2016). A language of play: New media’s possibility spaces. Computers and Composition, 40, 32‐47.
Ervin, Elizabeth. (1997). Encouraging civic participation among first-year writing students; or, why composition class should be more like a bowling team. Rhetoric Review 15(2), 382‐399.
Feigenbaum, Paul. (2012). Challenging rhetorics of adaptation through creative maladjustment. Composition Forum, 25.
Hooda, Saurabh. (2017, February 16). 3 reasons why everyone should learn programming. Entrepreneur India. Retrieved May 28, 2017, from https://www.entrepreneur.com/article/289248
Miller, Jessie; Gilkeson, Shanna; and Pignotti, Lisa. (2015, May 21). Hey, @students! #Letschat: Using social media to facilitate research and public engagement. PraxisWiki. Retrieved October 31, 2017 from http://praxis.technorhetoric.net/tiki-index.php?page=PraxisWiki%3A_%3ASocial+Media+for+Digital+Engagement
McFarland, Dave. (2014, November 20). The real reason why everyone should learn to code. Treehouse. Retrieved May 28, 2017, from http://blog.teamtreehouse.com/havent-started-programming-yet
O’Brien, Sara A. (2016, March 16). One tweet ruined her life. CNN Money. Retrieved May 28, 2017, from http://money.cnn.com/2016/03/16/technology/syfy-the-internet-ruined-my-life-gamergate-brianna-wu/
Ratto, Matt & Boler, Megan. (2014). Introduction. In Matt Ratto & Megan Boler (Eds.), DIY citizenship: Critical making and social media (pp. 1‐27). Cambridge, MA: MIT Press.
Sample, Mark. (2014, May 30). A protest bot is a bot so specific you can’t mistake it for bullshit: A call for bots of conviction.” Medium. Retrieved May 28, 2017, from https://medium.com/@samplereality/a-protest-bot-is-a-bot-so-specific-you-cant-mistake-it-for-bullshit-90fe10b7fbaa
Sharkov, Damien. (2017, November 3). Full list of Russian Twitter bots banned in election meddling probe. Newsweek. Retrieved November 27, 2017 from http://www.newsweek.com/full-list-russian-twitter-bots-banned-election-meddling-probe-700703
Shipka, Jody. (2011). Toward a composition made whole. Pittsburgh, PA: University of Pittsburgh Press.
Vee, Annette. (2013). Understanding computer programming as a literacy. Literacy in Composition Studies, 1(2), 42-64.
Wachs, Stephanie Wittels. (2016, April 20). The end of empathy. Medium. Retrieved May 28, 2017, from https://medium.com/@stephaniewittelswachs/the-end-of-empathy-5d8383b066d3#.gozsd213b
Whalen, Zach. (n. d.). How to make a Twitter bot with Google Spreadsheets (version 0.4) Blog post. Zach Whalen. Retrieved May 28, 2017, from http://www.zachwhalen.net/posts/how-to-make-a-twitter-bot-with-google-spreadsheets-version-04/
Whalen, Zach (2014, Nov 8). Protest bots. Open Transcripts. Retrieved November 27, 2017 from http://opentranscripts.org/transcript/protest-bots/
Welch, Nancy. (2005). Living room: Teaching public writing in a post-publicity era. College Composition and Communication 56(3), 470‐492.