BERKELEY, C.A. – At the University of California, Berkeley, two juniors studying computer science are battling fake news. “One of the things we wanted to see was where did this fake news originate from,” said Rohan Phadte. “How did it become so popular?”
Digging deep into Twitter, Ash Bhat and Rohan Phadte found many of the most angry and partisan tweets, on both sides, come not from real people but from automated Twitter accounts known as bots.
Using artificial intelligence they created a bot buster — “Botcheck.me” — that anyone can use to check a Twitter account.
“You can just go in and click that and in a few seconds, we get a classification,” said Bhat, as he demonstrated how the process works.
Then, “Botcheck.me” shows whether a tweet comes from a machine designed to spread fake news.
“These bots are like retweeting and amplifying voices in the Twitter community that otherwise would not be as amplified,” said Bhat.
One person can put a tweet out and then put together their army of bots that throw it out across Twitter.
“It seems like all Twitter is saying it right,” said Bhat. “It starts trending, you see hashtags like come up and it’s a few individuals that are able to really push this trend.”
Fake news stories about the 2016 election went viral, gaining readers and credibility — calling into question Twitter’s ability to monitor its platform. In a blog post, Twitter said it is battling the bots, catching “about 450,000 suspicious logins per day.”
But the students said their bot buster is still helping users discover thousands of bots on Twitter.
“Initially, this was just a project that we were like, ‘hey, this really annoys us,'” said Bhat. “Then all of a sudden we have thousands of daily active users that are using it every single day.”
A pair of college students might not win the war against fake news, but they have given those battling to defend the truth a new weapon.
© 2018 CBS Interactive Inc. All Rights Reserved.