|Iguala September 26, 2014, night from hell|
Pro-Government Twitter Bots Try to Hush Mexican Activists
More than 75,000 automated Twitter accounts (BOTS) are being used to combat protests and attack critics of the government, according to research presented by writer Erin Gallagher at the Chaos Communication Camp in Zehdenick.
The automated accounts, known as Peñabots, first appeared during the election of President Enrique Peña Nieto.
ON SEPTEMBER 26, 2014, a group of students departed the Ayotzinapa Rural Teachers’ College for a protest in Iguala, Mexico, about 80 miles away.(Note: actually they were going to DF to protest, not Iguala) They never arrived. What happened on the road to Iguala remains a mystery, but we know that at least three students were killed and another 43 are missing. The government’s official story is that the 43 students were killed after being handed over to the Guerreros Unidos cartel on the orders of the mayor of Iguala. But investigations conducted by the U.S. publication The Intercept paints a darker picture of complacency at higher levels of government.
The incident was emblematic of broader fears and frustrations with violence and corruption in Mexico and sparked a wave of ongoing protests across the country. Like so many modern protest movements, activists turned to social media to organize and promote their cause. One Twitter hashtag in particular—#YaMeCanse or “I am tired”—became a central hub for organizing protests and disseminating information.
That’s when artist and journalist Erin Gallagher, who covers the protests for Revolution News, noticed something strange. The search results for #YaMeCanse were flooded with tweets that included the hashtag but no other content, save for a few random characters such as commas, semicolans, and angle brackets. A typical tweet might be: “,,> #YaMeCanse.” The accounts tweeting the empty content bore the telltale signals of spambots, such as a lack of followers and a tendency to tweet variations of the same thing over and over again. It became difficult, if not impossible, for activists to actually share information with each other through the #YaMeCanse hashtag, and as a result it quickly dropped out of Twitter’s trending topics. Bots, it seemed, had effectively jammed the protesters’ communications channel.
Gallagher and LoQueSigue blogger Alberto Escorcia say the bots have followed protesters from hashtag to hashtag over the past few months, drowning out real conversations with noise. They’ve also seen similar bots create fake hashtags in apparent attempts to push real hashtags out of Twitter’s trending list, spread anti-protest messages, and even send death threats to specific activists.
It might seem petty to worry about Twitter bots in a country besieged with violence and corruption, but social media has become a central part of activism throughout the world. When it is undermined it has real effects. Yes, the term “Twitter revolution” is a massive simplification of the uprisings in Iran in 2009 or the “Arab Spring” protests that spread throughout the Middle East in 2010. After all, massive protests, riots and revolutions are as old as civilization itself.
But it’s fair to say that Twitter and other social networking sites have become the predominant platform for free expression throughout the world. And bots are now being used to stifle that freedom, not just in Mexico but around the globe. Similar tactics used against protesters in other countries, such as Turkey, Egypt and Syria. Just as the freedom of press was only truly guaranteed to those who could afford to buy one, freedom of social media may soon be limited to only those who can afford to build bot armies.
A Swarm of Lonely Bots
Escorcia has been watching the rise of Twitter bots in Mexico since the 2012 elections, when he noticed fake accounts promoting the candidacy of Enrique Peña Nieto, now the president. But the bots have become more active since the disappearance of the Ayotzinapa students.
Using social network visualization tools such as Flocker and Gephi, Escorcia has discovered a reliable way of detecting bot accounts by examining the number of connections a Twitter account has with other users. Bots have few connections, while real users tend to have far more. Using the software, he’s been able to identify many cases of bots used to sabotage protests.
Andres Monroy, a social-computing researcher at Microsoft Research, has observed similar behaviors in Mexico and other countries such as Venezuela. He says that in many cases, the bots are run by traditional spammers targetting popular hashtags — the kinds you might see hawking Viagra through the #FiveWordstoRuinADate hashtag — for commercial, as opposed to political, purposes. He also says that spam bots’ impact on trending topics is hard to quantify and that activists often mistake the natural ebb and flow of trends for malicious interference. “Just because a hashtag is popular doesn’t mean it’s trending,” he says. “Both sides [spammers and activists] put too much effort into trending topics.”
That said, Monroy confirms that politically motivated spam bots exist and pose real problems for activists, particularly when they flood hashtags with useless content. These efforts can have material effects on organizers. For example, Gallagher reports that as police dispersed a protest in Mexico City on December 1 of last year, protestors used the hashtag #RompeElMiedo (“break the fear”) to share information about police locations so that protesters, journalists and bystanders could exit protest without being arrested or beaten.
One protester posted a map highlighting an area to avoid in particular, but Gallagher says that Twitter accounts behaving in exactly the same way the #YaMeCanse spammer did made it extremely difficult for protesters to find that map. The result was that they were more likely to end-up being beaten or jailed.
The bots also make it more difficult for information to reach the public. For example, in February, activists used the hashtag #Acapulco to share photos from a protest in Acapulco, Mexico but the hashtag was pushed off the Twitter trending list by another hashtag, #SoyAmanteDe (“I am a lover of”), that Escorcia believes was created and promoted by bots, based on his analysis of the connections—or lack there of—between the bot accounts and other Twitter accounts. Tellingly, many of the #SoyAmanteDe Tweets used the same words, but in different arrangements, Gallagher reports.
Most chillingly, Gallagher says that activist Rossana Reguillo received death threats, including photos of charred bodies, over the span of two months, via Twitter bots.
So the question remains: who is running these anti-activist bots? Gallagher says that because of their appearance during President Peña Nieto’s campaign the bots are commonly referred to as “Peñabots,” and the ongoing assumption among activits on Twitter is that the federal government is behind the current crop of bots. Escorcia hasn’t been able to find any hard evidence linking the bots to the federal government in general, or to Peña Nieto in particular, but Gallagher points out that the bots seem to specifically target anti-government, as opposed to anti-cartel, sentiment on Twitter.
Below from Revolution News
This software can be modified to do more than just detect bots. Alberto Escorcia, lead developer from LoQueSigue, tells us that Gelphi has scientific origins and is used to analyze molecules, geophysical data and anything that creates networks and can be visualized. In the case of social networks it can be used to analyze the manner in which online conversation is carried out. For example during a natural disaster, conversation takes on a certain shape. Conversation about Christmas takes on a different geometric form. When bots attack they have a very particular shape. Before any event a certain pattern is created. If we have the ability to analyze all tweets in Gelphi, it would be possible to detect trending topics before they happen
Below from Vice
Below from Vice