There's a lot of bots out there on Twitter.
That's the message from a new Pew Research Center study, out Monday, which found that two-thirds of tweets that link to digital content are generated by bots — accounts powered by automated software, not real tweeters.
Researchers analyzed 1.2 million tweets from last summer (July 27-Sept. 11), most of which linked to more than 2,300 popular websites devoted to sports, celebrities, news, business and sites created by organizations.
Two-thirds (66%) of those tweets were posted or shared by bots and even more, 89%, of links that led to aggregation sites that compile stores posted online were posted by bots, the study says.
The findings suggest that bots "play a prominent and pervasive role in the social media environment,” said Aaron Smith, associate research director at Pew, which used a "Botometer" developed at the University of Southern California and Indiana University to analyze links and determine if was posted by an automated account.
“Automated accounts are far from a niche phenomenon: They share a significant portion of tweeted links to even the most prominent and mainstream publications and online outlets," Smith said in comments accompanying the study. "Since these accounts can impact the information people see on social media, it is important to have a sense their overall prevalence on social media.”
The Pew researchers did not attempt to assess the accuracy of the material shared by the bots. Also not determined: whether the bots were “good” or “bad" or "whether the content shared by automated accounts is truthful information or not, or the extent to which users interact with content shared by suspected bots,” said Stefan Wojcik, a computational social scientist said in the study.
Other findings about bots:
•Bots were responsible for about 90% of all tweeted links to popular adult content websites, 76% of popular sports sites, and 66% to news and current events sites.
•Some bots do more work than others. The 500 most-active suspected bot accounts sent 22% of tweeted links to popular news and current events sites. In comparison, the 500 most-active human tweeters sent about 6% of links to the outlets.
•Bot accounts with a political bias were equally liberal or conservative.
Bots have long plagued Twitter and other researchers have estimated as many as 15% of all Twitter accounts could be fake. Twitter has said the number is lower.
Twitter's rules allow automated software, but bans the posting of misleading or abusive content or spam. In February, Twitter suspended multiple accounts following the indictment by special counsel Robert Mueller of Russian nationals for meddling in the U.S. election, including using fake Twitter accounts to wage "information warfare" against the U.S.
The social network also attempts to remove deliberately manipulative tweets and shared offered details about that process Friday in an online post from Del Harvey, Twitter's vice president for trust and safety.
During Tuesday's shooting at YouTube headquarters in San Bruno, Calif., Twitter began to see accounts "deliberately sharing deceptive, malicious information, sometimes organizing on other networks to do so," Harvey says. That activity is typical as information about tragedies emerges and presents "an especially difficult and volatile challenge" in how to respond to "people who are deliberately manipulating the conversation on Twitter in the immediate aftermath of tragedies like this," she says.
Twitter "should not be the arbiter of truth," Harvey says, but the network does have rules against abusive behavior, hateful conduct, violent threats, spam and against suspended users creating new accounts.
In recent months, Twitter has improved its tools and ability to respond to manipulative activity on the service, she says. After the YouTube shooting, "we immediately started requiring account owners to remove Tweets — many within minutes of their initial creation — for violating our policies on abusive behavior," she says. "We also suspended hundreds of accounts for harassing others or purposely manipulating conversations about the event."
Automated systems also helped prevent suspended tweeters from creating new accounts and helped find "potentially violating Tweets and accounts" for the staff to review, she says.
At the same time, the Twitter team "was also focused on identifying and surfacing relevant and credible content people could trust," Harvey says. "Moments highlighting reliable information were available in 16 countries and in five different languages — many within 10 minutes of the first Tweets — and also surfaced within top trends related to the situation."
Twitter continues to deploy technology and people to improve the situation, she says. "We're committed to continuing to improve and to holding ourselves accountable as we work to make Twitter better for everyone," Harvey says. "We’re looking forward to sharing more soon."
