Robots are better than humans at fomenting human interaction on the Internet and creating online communities, according to recent experiments in social engineering.
The research was conducted by the Web Ecology Project, a community of researchers dedicated to better understanding Web culture through quantitative research.
“We wanted to see if it’s possible to design bots that have a statistically significant and reliable effect on the way people are connecting and talking online,” said Tim Hwang, WEP’s director and an organizer of the experiment.
During the two-week experiment, Socialbots 2011, hundreds of targeted Twitter users unwittingly conversed with the tweets of a robot.
Three teams had designed software to bring forth as many human responses and interactions as possible.
“People always say that they’re great at building communities online, but people are limited in some ways,” Hwang said. For example, “they only have so many hours in a day.”
The winning "socialbot," JamesMTitus, kept interchanges going for a record of 12 consecutive tweets and garnered close to 200 overall responses from targeted Twitter users in less than a week. Humans who had created new Twitter accounts in a similar experiment garnered at most 60 responses, Hwang said.
To achieve this, Titus, a bot with a “very promiscuous, yet lovable persona” and who is “obsessed with his pet cat,” according to the blog of one of its creators, tweeted a generic question to one of the Twitter users being targeted about every seven minutes.
If the human responded, the robot replied with its own generic response. For example, when Titus asked one user what his favorite Dr. Seuss book was, the person responded, “the older cartoon of the Grinch, but not the Jim Carrey version.” “Rock on,” replied the robot.
But Titus did more than fool people into revealing their views on Dr. Seuss to a robot, and this is what truly intrigued the researchers at WEP. Not only did Titus’s tweets elicit responses, they also served as a bridge between individuals and communities. Twitter users unknown to one another began to interact because of the robot’s tweets, and in that sense, Titus had a more lasting and valuable effect on the Twitter landscape.
“Bots have been trying to get you to do something minor like click on a link or buy Viagra for years, so the coding is not too novel,” Hwang said. “But we’re trying to get them to talk to someone who will then talk to someone else.”
Hwang’s team is now working to get networking bots to stimulate more than random interaction between individuals and mobilize them in a more targeted way. Rather than just getting people to exchange trivial tweets, they’d be rallying behind a cause.
“My dream is to make bots targeted agents to connect particular groups,” Hwang said. “If we want people to be more civically engaged, for example, we can try to build a bot that would achieve this. In the future, you can design the kind of community that you need for a cause. Or a bot could be used to inject a news story into communities and thus control the news cycle.”
If all of this sound too cunning for comfort, Hwang retorts that this is no different than advertising.
“If you’re opposed to something like bots, you should be opposed to advertising. It’s using mass communication as a (form) of persuasion. It could be of great use,” he said.
“Certain people are great connectors, but they’re few and far between. Maybe bots can provide a synthetic version of this and make cool things happen that wouldn’t otherwise happen. They can create the social scaffold that allows people to do things that they would have wanted to do anyway, if only they’d known the other community existed.”