Microsoft research lent a hand, too.
A major tech corporation's research wing teams up with a group of Mexican communist revolutionaries to make bots that fight government corruption. It sounds like the plot of a deliciously cheeky 80s sci-fi B-movie, but I'm here to tell you that this really happened in 2015.
A paper published to the ArXiv preprint server this week describes how a Microsoft researcher worked with computer scientists from West Virginia University and UC Santa Barbara to create a Twitter platform called "Botivist," which acts as a tool for activists to recruit volunteers on social media for campaigns. It identifies likely candidates from their past tweets—if they recently used the keyword "corruption," for example—and sends them targeted calls to action using Twitter accounts run by bots.
"Citizens often don't contribute to their community because they don't have time," Saiph Savage, assistant professor at West Virginia University and lead author on the paper, told me. "We can imagine a world where we have agents that guide us—where, as a citizen, can you make the best contribution to people in your city?"
"We use and build technology to help the people"
Botivist's first test was a campaign to raise awareness about governmental corruption with input from members of the Mexican Grupo de Accion Revolucionaria [link in Spanish], or the Mexican Revolutionary Action Group (GAR). The GAR is openly and radically communist, and aligns itself with Trotskyism, which holds that all the world's workers must march into their glorious socialist future together. They also fight against corruption and for worker's rights in Mexico.
Some of its members also happen to be engineering students at the National Autonomous University of Mexico (UNAM), where Savage was based during the work on Botivist.
"We are planning to use Botivist to recruit technology experts who are also interested in doing social good in their cities or country," Walter Angel, a UNAM student and leader of the GAR, wrote me in an email, which was translated by Savage. "We have several projects where we are empowering rural communities in Mexico like giving them more bio-friendly stoves."
In August, security researchers presented work alleging that more than 75,000 Twitter bots in Mexico are currently working to spread political propaganda and silence dissent. In one case, a hashtag warning protesters to stay out of a particular area was spammed by the bots, and shortly after police began arresting and beating protesters. Botivist, then, can be seen as a way of using those same tools for good instead of harm.
As for Microsoft's involvement, the paper appears on the company's research site. Andres Monroy-Hernandez, the Microsoft researcher, told me that the project was part of his own research agenda, which every Microsoft researcher is permitted to explore. The paper will be presented at the 2016 ACM conference on Computer-Supported Cooperative Work and Social Computing, according to Microsoft's site.
In the test with GAR, Botivist sent out 376 calls to action requesting ideas on how to fight corruption to Twitter users split into four categories. The first category was targeted with messages simply asking to brainstorm ideas. The other three groups were sent messages paired with different emotional appeals like, "Remember that: One for all, and all for one!" With the more direct approach, 81 percent of users responded to the question asked by the bot, and even to follow up questions.
Interestingly, the authors note, the messages that tried to persuade volunteers sometimes resulted in circular discussions about the value of bots in activism. Apparently, when a robot is trying to convince you of something, it's a little more suspect than when it is clearly being used as an administrative tool by people to drive social change.
"It is our social responsibility to create change that will empower the people," Angel wrote. "We use and build technology to help the people."
UPDATE: The Microsoft researcher contributing to the work on Botivist responded to Motherboard's request for comment, and this article has been updated with their comments.