Bots are one of the best parts of Twitter. If you’ve spent much time exploring the sprawling social-media platform, chances are you’ve followed at least a few of them. You might’ve followed @tinycarebot, for example, which periodically reminds you to breathe, go outside, or take a nap.
You might’ve followed @netflix_bot, which provides you with a steady stream of shows that have just become available on Netflix.
Or maybe you followed @museumbot, which fills your Twitter feed with images from the Metropolitan Museum of Art’s online archive.
If these or any other non-spammy bot accounts have brought you automated joy as you’ve trawled through Twitter’s often hostile landscape, we have bad news: Some of your favorite bots are about to go silent.
Many of these delightful and creative accounts will disappear in the coming months due to a company-wide attempt to eradicate malicious bots from the platform. Though this is a well-intentioned effort to curb computational propaganda, it will likely sweep up art bots in its wake.
Until now, Twitter was a bot-maker’s dream. It provided access to the platform’s data through open and robust APIs, which are back-end interfaces that allow bots to receive data. But in July, Twitter announced that they would now require bot developers to undergo a comprehensive vetting process before they can gain API access. This means that casual bot-makers will have to request a formal developer account, which involves providing detailed information about how they plan on using Twitter’s data streams to make automated art.
For Allison Parrish, who has been making bots on Twitter since 2007, this change in policy contradicts the spirit of bot-making, which she considers an experimental, DIY, almost subversive art form.
Parrish’s first bot was called @everyword, and it did nothing but tweet every word in the English language in alphabetical order.
Since then, she has made dozens more, but her favorite is @the_ephemerides, which tweets out raw images of distant planets from NASA’s archive, coupled with computer-generated poetry.
When Parrish first started making experimental bots back in 2007, it was unchartered territory. “I think @everyword was one of the earliest bots on Twitter made specifically for trouble-making, artistic purposes,” she said in an interview with Quartz. But she soon became part of a vibrant community of bot-makers who also developed weird, poetic, satirical, funny, and informative bots for the platform.
Parrish says Twitter’s policy could destroy the creative spirit of this community. “Asking permission to make a bot is like asking someone permission to do graffiti on a wall,” she says. “It undermines everything that is interesting about bot-making.”
Parrish’s graffiti analogy is apt. Twitter’s new developer policy is part of a broader attempt to rid the platform of spam and malicious bots, making it a cleaner, more sanitized place to spend time. The question, though, is whether this digital gentrification will sacrifice the very essence of what made Twitter a compelling and creative place to begin with.
I’ve been following creative bot accounts for years. They make my Twitter feed weirder and funnier, a place of ontological ambiguity where tweets from journalists and politicians are interspersed with moments of random, computational beauty.
Some of my favorite bots include Everest Pipkin’s @tiny_star_field, which tweets out random constellation of dots and asterisks; Chris and Ali Rodley’s @MagicRealismBot, which generates premises for fantastical narratives; and Jia Zhang’s @censusAmericans, which writes brief, poignant biographies of anonymous Americans by compiling information from open census data.
But while Twitter’s lax automation policies have encouraged creative bot-making, it has also made it a platform teeming with malicious bots. These include spambots that aggressively market malware, troll bots that provoke angry debate, fake human accounts that artificially raise followers, and misinformation bots designed specifically to undermine public discourse and promote hate.
Over the years, the number of these malicious bots has grown, as have their influence on real-world events. As well as interfering in the US presidential election, Twitter bots have been used to manipulate stock markets, steal identities, and harass and threaten minorities. This destructive automated activity has ultimately decreased Twitter’s signal to noise ratio and eroded trust on the platform.
Twitter has tried to deal with this issue in the past by using anti-spam filters to identify bot activity and eradicate the associated accounts, but up until this point, these efforts have been in vain. Earlier this year, however, the platform stepped up its anti-bot program by creating restrictions around who can have access to the platform’s APIs. They’re hoping this extra step will help stymie the flow of malicious bot accounts.
These new rules do not only apply to new developers, but anyone with existing developer access. This means that if you already have a bot on Twitter, you have to justify its ongoing existence by retrospectively explaining what it does and what purpose it serves on the platform.
Twitter maintains that introducing this review process will allow them to catch malicious accounts before they gain access to the platform. It acknowledges that this will “add an extra step and time to get started with development,” but believes that this is a sacrifice that must be made to “help increase the collective health, openness and civility of public conversations,” as Jack Dorsey puts it.
While many of Twitter’s creative bot-makers are undertaking the administrative task of making their bots comply with these new rules, others are choosing not to.
Darius Kazemi, who has been called “the Oscar Wilde of internet bots,” says that it would simply be too much work to update his 80+ bots on the platform, which include @TwoHeadlines, @rapnamebot, and @expandingbrainbot. “It is prohibitive for me to justify their existence,” he said in an interview with Quartz.
Parrish has also said that she will not be taking any “proactive measures” to keep her bots working on Twitter. This means that in the next few months, their bots, which are some of the most beloved and long-standing on the platform, will probably begin disappearing from our feeds. When Quartz asked the bot-makers whether any had already been banned, both Parrish and Kazemi said that they had stopped checking.
For these bot-makers, letting their creations die off on Twitter is an act of protest. It’s not so much directed at the new developer rules, but at the platform’s broader ideology. “For me it’s becoming clear that Twitter is driven by a kind of metrics mindset that is antithetical to quality communication,” Parrish says. “These recent changes have nothing to do with limiting violent or racist language on the platform and are all about making it more financially viable.”
Kazemi agrees, adding that to continue making creative bots on Twitter is making a bargain with the devil. “We’re being asked to trade in our creative freedom for exposure to a large audience,” he says. “But I am beginning to suspect that once we all leave Twitter, they will realize that we represent a lot of what made Twitter good, and that maybe the platform needs fun bot makers more than we need Twitter.”