Skip to Content

How the Bot-y Politic Influenced This Election

Nearly 20 percent of all election-related tweets come from an army of influential robots.
November 8, 2016

If your political conversations on social media seem mechanical and predictable, it might be because you are debating with a robot.

A study published the day before the election found an estimated 400,000 bots operating on Twitter that were tweeting—and being retweeted—at a remarkable pace, generating nearly 20 percent of all election-related messages.

Emilio Ferrara

Besides being numerous, these bots are also quite influential, and capable of distorting the online debate, according to authors Alessandro Bessi and Emilio Ferrara of the University of Southern California Information Sciences Institute.

One thing remains mysterious: who is creating them? That’s still impossible to determine, Ferrara told MIT Technology Review in a discussion about his study, conducted over a month this fall, a period that included all three presidential debates. Below is an edited transcript.

In the study you measure the influence of bots, and find it to be surprisingly high. Can you explain that?

We measure how many connections an account has to other accounts and how many different users retweet that account.  What we find is that people are not really good at determining whether the information sources they are exposed to are a human or a bot. Bots are being retweeted at the same rate as humans. Since bots are much more active and produce [many more tweets] the by-product is they are being retweeted a lot and becoming influential. That leads to some serious issues with retweeting—spreading misinformation, spreading rumors, conspiracy theories, and so forth.

This map from the study shows the volume of bot tweets generated between September 16 and October 21, 2016 by state, indicating that one state, Georgia, is the biggest contributor.
This map from the study reflects the volume of human tweets generated between September 16 and October 21, 2016 by state. It shows the conversation being driven by the most populous states, including California, Texas, Florida, Illinois, and New York.

Do you expect this to have an impact on the outcome of the election?

It is really hard or maybe even impossible to make an interpretation of what influence these dynamics have on the outcome of the election, the actual voting and so forth. One thing at risk is voter turnout. We find in some states, in particular in the South and the Midwest, there are way more bots than in any other state. People might think there is a real grassroots support there, but in reality it’s all generated by the bots.

Nearly 75 percent of the bots you found were supportive of the Republican candidate Donald Trump, and the messages in their tweets were different, too.

The bots supporting Donald Trump are really producing an overwhelmingly large amount of positive support for the candidate, while in the case of Hillary Clinton supporters, more tweets are neutral than positive.

Social media bots getting involved in politics isn’t a totally new phenomenon, but you find that this year’s batch is particularly sophisticated.

These bots are more complex, using artificial intelligence to chat with people. They can aggregate the sentiment in a polarized discussion and maybe even further polarize it.

It had been easier to identify earlier bots, but now it’s incredibly difficult for a human to make a determination. I did a test on myself, and in some accounts there are signs that are clear, such as posting 1,000 tweets per hour. In others it’s more difficult. They look like they go to sleep [going offline for a prolonged period each day]. They tweet five, 10, 15 tweets in a row, and then none for an hour. They clone the behavior of people.

Keep Reading

Most Popular

It’s time to retire the term “user”

The proliferation of AI means we need a new word.

The problem with plug-in hybrids? Their drivers.

Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.

Sam Altman says helpful agents are poised to become AI’s killer function

Open AI’s CEO says we won’t need new hardware or lots more training data to get there.

An AI startup made a hyperrealistic deepfake of me that’s so good it’s scary

Synthesia's new technology is impressive but raises big questions about a world where we increasingly can’t tell what’s real.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.