Select your localized edition:

Close ×

More Ways to Connect

Discover one of our 28 local entrepreneurial communities »

Be the first to know as we launch in new countries and markets around the globe.

Interested in bringing MIT Technology Review to your local market?

MIT Technology ReviewMIT Technology Review - logo

 

Unsupported browser: Your browser does not meet modern web standards. See how it scores »

{ action.text }

 

Sometimes fascinating discoveries are made entirely by accident. This is a good example. A few years ago, Luca Maria Aiello and a few pals from the University of Turin in Italy began studying a social network called aNobii.com in which people exchange information and opinions about the books they love. Each person has a site that anybody can visit. Users can then choose to set up social links with others.

To map out the structure of the network, Aiello and co-created an automated crawler that starts by visiting one person’s profile on the network and then all of the people that connect to this node in turn. It then visits each of the people that link to these nodes and so on. In this way, the bot builds up a map of the network.

Crucially, to gain access to the network, the team had to create an empty user account for their crawler which they called lajello.

The team let the lajello crawler loose in September 2009. At this time the network was small enough for lajello to map out the entire structure once every 15 days or so.

Then in July 2010, aNobii.com changed its default user settings so that every user could see all the others that had visited their personal site. “As a result, our crawler left a trace of its passage in all the profiles reached approximately twice a month,” say Aiello and co.

And curiously, people began to respond to the crawler’s visits. That gave the team an idea. “The unexpected reactions the bot caused by its visits motivated us to set up a social experiment in two parts to answer the question: can an individual with no trust gain popularity and influence?” they say.

Aiello and co were careful to ensure that the crawler did not engage with anybody on the network in any way other than to visit his or her node. Their idea was to isolate a single, minimal social activity and test how effective it was in gaining popularity.

They began to record the reactions to lajello’s visits including the number of messages it received, their content, the links it received and how they varied over time and so on.

The results surprised them. Every time lajello began its round of visits, it triggered a burst of comments on its public wall. When it finished its round, people quickly stopped sending messages but resumed at the same intensity when the bot started visiting again.

By December 2011, lajello’s profile had become one of the most popular on the entire social network. It had received more than 66,000 visits as well as 2435 messages from more than 1200 different people.  In terms of the number of different message received, a well-known writer was the most popular on this network but lajello was second.

“Our experiment gives strong support to the thesis that popularity can be gained just with continuous “social probing”,” conclude Aiello and co. “We have shown that a very simple spambot can attract great interest even without emulating any aspects of typical human behavior.”

But this was just the start of their experiment. Having created all this popularity, Aiello and co wanted to find out how influential the spam bot could be. So they started using the bot to send recommendations to users on who else to connect to.

The spam bot could either make a recommendation chosen at random or one that was carefully selected by a recommendation engine. It then made its recommendations to users that had already linked to lajello and to other users chosen at random.

Again, the results were eye-opening. “Among the 361 users who created at least one social connection in the 36 hours after the recommendation, 52 per cent followed suggestion given by the bot,” they say.

But the targeted recommendations given to followers were far more effective than those given to non-followers. “In other words, lajello has a greater persuasive power over those who are more aware of its presence and activity,” say Aiello and co.

Incidentally, Aiello and co had to terminate their experiment when aNobii.com suspended lajello’s account following widespread discussion on the network over whether lajello was human or not.

Nevertheless, that’s interesting work that shows just how easy it is for an automated bot to play a significant role in a social network. Popularity appears easy to buy using nothing more than page visits, at least in this experiment. What is more, this popularity can be easily translated into influence.

It is not hard to see the significance of this work. Social bots are a fact of life on almost every social network and many have become so sophisticated they are hard to distinguish from humans. If the simplest of bots created by Aiello and co can have this kind of impact, it is anybody’s guess how more advanced bots could influence everything from movie reviews and Wikipedia entries to stock prices and presidential elections.

Ref: arxiv.org/abs/1407.8134 : People are Strange when you’re a Stranger: Impact and Influence of Bots on Social Networks

5 comments. Share your thoughts »

Tagged: Web, Mobile

Reprints and Permissions | Send feedback to the editor

From the Archives

Close

Introducing MIT Technology Review Insider.

Already a Magazine subscriber?

You're automatically an Insider. It's easy to activate or upgrade your account.

Activate Your Account

Become an Insider

It's the new way to subscribe. Get even more of the tech news, research, and discoveries you crave.

Sign Up

Learn More

Find out why MIT Technology Review Insider is for you and explore your options.

Show Me