What is a bot?
The modern world is driven by the internet, especially social media. The popular microblogging site, Twitter, claims to be “your window to the world”, with several hundred million active users posting millions of tweets per day. Bots, little bits of code that do a thing, are everywhere, especially posting on Twitter.There is even a ‘botifesto’ extolling the virtues and possibilities of bots, not just those on Twitter, and the myriad actions that they are designed to do. It tries to capture the full width and breadth of what bots are and what they could be.
On Twitter with its set 140 character limit on posts and expectations on what those posts should look like, where artists and programmers have turned those bits of code into a new form of internet-based art, it is easier to create a bot that does something interesting or different than it would be anywhere else on the web.
But, what is a Twitter bot? The most apt definition I could find was from The New Yorker:
“Twitter bots represent an open-access laboratory for creative programming, where good techniques can be adapted and bad ones can form the compost for newer, better ideas. At a time when even our most glancing online activities are processed into marketing by for-profit bots in the shadows, Twitter bots foreground the influence of automation on modern life, and they demystify it somewhat in the process.”
It mentions several reasons why someone would want to preserve a bot: to study or learn from its code, to understand what it says about modern culture and modern life. But I would add another reason: simply because they find it funny. There are already researchers studying what the bots say about modern culture, either through their posts or through those that interact with the bot.
Why this bot?
Two Headlines takes two of the news headlines from Google News and then posts that combined result. The posts give a humorous, if slightly jumbled, look at the current, important event happening around the world, at least according to Google. In under three years, the bot has managed to post more than 20,000 times and gain over 5,000 followers. While not an internet-high, it is a respectable following for something that is not advertised, instead relying solely on word of mouth.
The creator of the bot once described Two Headlines by saying,
“Part of the reason it’s funny is it’s timely — it’s always talking about what’s in the news right now because it’s pulling from Google News. The other advantage is that, much like Twitter, news headlines have a very specific way they’re written, both within publications and across publications. … It plays with the convention of headline-writing itself and subverts those expectations. Its hit rate is very high. Probably four or five tweets a day are very funny, which is a pretty high hit rate for a bot.”
Programs and their code are always studied by other programmers and those wanting to become programmers. People will always want to know how things work. Two Headlines’ code is freely available online and it has already been commented to help and explain the parts of the program. The comments were designed to allow others to modify the program for different results, which would make understanding the program easier for those with little to no programming experience. It is already being used as a teaching tool for those that want to learn about bot creation.
Preserving the code would be valuable to people that are interested in studying programming and/or Twitter bots or those that are studying online culture. However, there is no mention of if there have been revisions to the code, so there would be no way to preserve older versions of it, if they exist without the help of the bot’s creator.
Who said what?
There is also the context and commentary surrounding twitter bots that would be useful to anyone studying bots, especially those looking at them as more than fancy bits of programming. There have been many articles written about bots, their creators, and their cultural effects, not counting the articles trying to find the most interesting bots to follow. While Two Headlines may not have gotten that much press specifically dedicated to it, or any bots dedicated to mocking or adding to it as other popular bots have, at least none that I’ve found, it is still mentioned in the media, just not as often as its creator.
Speaking of the creator, let us not forget that Two Headlines is a program and, therefore, created by a person, in this case by Darius Kazemi, who is a prolific bot creator. It says so on his twitter and website, complete with links to the other projects he is working on or has created. There is also a bio and links to several news stories on the website.
In addition to his bot creations, Kazemi has done a lot of work to help others make their own bots and is responsible for Bot Summit, a conference “where botmakers from around the world get together, both in person and online, to discuss the art and craft of making software bots”. In doing all this bot-related work, he has developed a following of fans from many different fields, such as other programmers, game developers, comedians, philosophers and even an English literature professor, at least according to one article. Ian Bogost, whom you might remember being mentioned a few weeks ago on this blog, was quoted as saying, “You have a favorite comedian or favorite artist and you look forward to what they say, because you want to see the world through their eyes. The same kind of thing is happening with Darius.”
I think there’s a lot to be said for the bot as a means of social commentary. There is a bot that tracks any time a Wikipedia entry is changed by a computer with a Congressional IP address (it was designed by an UMD student if I remember correctly). While it mightn’t be uncovering the depths of political spin, I think it doesn’t hurt to show how deeply bots can dig into factual information, even if they’re being funny. What do the analytics from Two Headlines show about our news obsessions? How bad has headline writing (possibly intentionally) become?
For context on this, that bot is Congress Edits and Ed Summers made it.
You do a nice job laying out the case for bots in general and you lay out a good bit on why Two Headlines in particular is significant.
You give us a bit of insight into some of the elements of the work and who they might be significant to, but it would be great to get into some more detail there. That is, what elements of Two Headlines would be most relevant to preserve or document to meet the needs of potential different current and future stakeholders. You’ve got the source code for the bot, the corpus of tweets it has produced, reactions to it, the actual experience of it working in real time (which is contingent on the Google News API, and the various platforms it requires to run). Aside form that, there is also a case to be made that the bot’s significance as a teaching tool for others to look at for making bots might suggest interviewing various bot makers to document what they find compelling or important about it, how they heard about it, what they think was important about it when it launched etc.
Depending on where you go with your approach in your preservation intent statement, it will likely be important to get more into want to what exactly Two Headlines the program runs on. I think you can use the Bogost and Montfort book as a way to think about the platforms that it runs on. You can look through the explanations in the source code and the dependent libraries and systems that it runs on in it’s GitHub repository. https://github.com/dariusk/twoheadlines So for instance, it runs on Twitter and uses Twitter’s API, it requires Node.js and npm.js. Given that, it would be worth thinking through the extent to which those dependencies are things that one would want to save along with the source code.
This is all to say; I think you are off to a great start. You’ve met the objectives of the assignment and I am looking forward to seeing how this develops in your preservation intent statement.
I was trying to find posts about people’s experiences with the bot or its code, but there is very little that I could find. Hopefully, I can find someone to talk with about the bot, I would really like to have someone’s first-hand knowledge about what makes Two Headlines significant. I did try to contact the developer to get his input on the endeavor, but have yet to hear anything from him.
As I have little programming experience and none with APIs and what it takes to make something that posts on Twitter, the more technical aspects of this are definitely something that I’m going to have to learn a lot more about before this project is over. As I plan on preserving the source code, the platform elements and dependencies will need to be explored to at least determine which are vital to the understanding and functions of the bot and if its necessary to preserve them along with the bot, or if a copy preserved elsewhere will be sufficient.
Especially timely, considering Microsoft’s recent failed attempt at bot-making. This article from Vice talks about the reactions within the botmaking community to Microsoft and Tay, including several statements by Kazemi. He mentions Two Headlines in particular, and tweaks he made to the code to try and prevent gender mismatches between the two headlines, to cut down on transphobic jokes, etc. Word blacklists are another thing it covers, for racism, sexism, etc, and generally it looks like there are a few useful links to other creators as well.
Interesting article, I never knew that twitter bots were that complex and had that kind of value. One thing to consider when deciding how to preserve them is ownership. Do bots count as property? What about what they collect? I am interested in finding out.
I really like the definition of “bot” that you included here. It captures the weirdness of bots being abruptly recontextualized (from private to public, from infrastructure to art) and the general public excitement around bot-building. And also (bonus!) lends itself to writing about “promise and peril” and other thinkpiece-type headline snippets. Which could then be remixed into @TwoHeadlines tweets. Actually, maybe the strangeness is probably what makes bots funny to people: classic first reaction when one doesn’t know what to do. People tend to focus on the beautiful when talking about my bot (@mothgenerator) and tend to overlook some of the textual hilarity, but it’s there.
Looking forward to seeing how you document @TwoHeadlines as a Darius Kazemi joint (sorry). With the wealth of bot-related material he’s produced, it’s going to be fun narrowing down a scope for this project.