There’s a new online army of internet trolls, and instead of trying to hack networks they’re simply trying to create chaos with false reports of breaking news.
Many online trolls act independently, posting inflammatory content online to get a rise out of others. But a New York Times report details an entire Russian agency whose goal is promulgating misinformation and even creating complete hoaxes.
One such hoax happened late last year when a series of random Twitter accounts began tweeting about a disaster at a Louisiana chemicals plant. Using the hashtag #ColumbianChemicals, tweets began bombarding media outlets and politicians about a supposed explosion at the plant. It turns out that no such disaster actually happened, and according to The New York Times this misinformation campaign likely stemmed from a Russian agency that works day in and day out to spread both national and international misinformation.
Here’s what one tweet about the fake explosion looked like:
But is it possible to know to recognise the fake news trolling from the real breaking news stories? According to the cyberthreat analytics company Recorded Future — whose software has been shown to predict the future — it is.
The Cambridge-based company analysed the social media data of the #ColumbianChemicals debacle and found three characteristics that differentiated this particular hoax from the real disasters being reported on social media.
1) When disaster strikes and people are tweeting about it, there is a “vocal minority” that tweets louder and more often
Recorded Future explains that this means there is an extremely small number of people who tweet more frequently than the rest. This would likely be a few people at the scene of an incident tweeting rapidly what is happening in real-time, or those who are “extremely upset about the incident.”
In most cases, the number of “vocal minorities” is extremely disproportionate from the rest. For #ColumbianChemicals, this was not the case.
“Instead, we see suspiciously smooth patterns in the data. It doesn’t look lumpy like real data – it looks very produced,” Recorded Future writes.
Here are two tweet graphs — one from #ColumbianChemicals and another from when the real computer vulnerability called #Sandworm was first discovered:
Notice how the #ColumbianChemicals graph decreases at a constant rate. This means that there was a suspiciously high number of twitter accounts tweeting more than once about the event. Compare this to #Sandworm, where a very small number of people tweeted numerous times, but the vast majority of Twitter users only tweeted about the event once.
2) The accounts being targeted by the #ColumbianChemical tweets didn’t make sense
Usually, when disaster strikes, people are trying to get the attention of the authorities to fix the problem. The accounts sending these tweets “fanned out” who they were targeting. And the people who received the most tweets were national news outlets and politicians — none of the tweets tagged accounts of people appropriate to the situation.
Here’s a list of the most targeted Twitter accounts from #ColumbianChemicals.
3) When people begin tweeting about breaking news, they quickly gain an audience
Citizen reporters who were at the heart of Ferguson quickly gained a captive audience of Twitter users looking for up-to-the minute updates. When the #Sandworm vulnerability was first disclosed, those tweeting their knowledge of malware quickly attracted a new following. For #ColumbianChemicals this was not the case. In fact, most accounts tweeting about the situation accrued zero new followers. This was a huge red flag.
Recorded Future was able to use its tools to find these anomalies, all of which help explain why #ColumbianChemicals was a hoax, and how to recognise similar cases.
As trolling campaign become more widespread, firms like Recorded Future hope to use these sorts of tools in real-time to weed out the real breaking stories from the fake.