Ben Nimmo Profile picture
Writer, linguist, diver. Principal Investigator, Intelligence & Investigations @OpenAI. RT ≠ endorsement.

Apr 15, 2018, 18 tweets

Thread: A few thoughts on the Pentagon's claim of a "2000% increase in Russian trolls."

First point: it takes quite a lot of evidence to identify a genuine Kremlin troll. Examples in this post:

medium.com/dfrlab/trolltr…

I had a look at traffic on the @DeptofDefense handle from 0800 ET on April 13 to 0800 ET on April 14, using @Sysomos.

Note the surge around 2100, when the strikes were announced. It's about a 1400% increase in traffic.

But most of the traffic was driven by retweets of the Pentagon's own posts. Of the 10 most retweeted posts, 6 were from DoD.

Only one of the top 10 came from a Russian account: the Russian embassy in the USA.

It can be troll-like, but it's hardly a covert operation, or a major surge.

The DoD Facebook page got a fair amount of traffic, but it was mostly positive. Again, not what you'd call a troll surge.

I had a look at traffic on #SyriaStrikes, too. Much higher volumes here: 300k posts, mostly tweets, in 24 hours.

That's not unusual for a breaking topic. "Covfefe" racked up 100k tweets in the first 45 minutes alone.

A lot more of the traffic here was negative. Sentiment analyses should always be read with caution, but an eyeball scan confirmed the trend.

One stand-out post straight away: this one had 17k retweets, from an account which only has 341 followers.

That's weird, but an initial scan showed human-looking amplifiers, not bots.

Anyone got time for a deep dive? @AAlaphilippe @lmneudert @katestarbird

Overall, though, the main drivers of traffic were verified accounts with their own perspectives and their own followings.

Could be some bot amplification in the mix, but these are high-influence users. And not "Russian trolls."

There were some genuine pro-Kremlin trolls in the mix, and they were influential, but these are old friends, and this is their business as usual. Not a surge.

I took a look at activity around @StateDept, too. This tweet looked like a likely troll magnet.

Some of the replies did, indeed, look a bit trollesque.

Note the spelling "desinfo" here.

This account shares a lot of Kremlin-style narratives.

Its English is reminiscent of known Russian-language trolls, too.

"For which reason Russia will down a civil plane?"

It shares a lot of Russian and Iranian content. Three of these shares are direct copies of tweets from Slutnik, one each from Mehr News and Press TV.

Sorry, +Sputnik+. Though the other version could be seen as fitting.

Anyway, it shares the occasional Russian-language content, too. It's definitely pro-Kremlin. Part of an organised campaign, or a soloist? Unclear.

@conspirator0 @propornot

This one got annoyed at State, too. It shares a lot of Kremlin-style messaging on some key themes, such as Ukraine, Crimea and the White Helmets.

Curiously, it never posted about Turkey until November 2015, when Turkey downed a Russian jet. Then it suddenly became very interested in the Turkey-ISIS-oil trade story.

But NB, it doesn't seem to have posted on MH17 or Skripal. Pro-Kremlin, yes. Kremlin-linked? Questionable.

Upsum: There was a lot of activity on #SyriaStrikes, and a surge around the DoD handle.

The usual pro-Kremlin accounts, and some possible extras, in the mix.

But a 2000% increase? Not based on these scans.

"Russian trolls". Use the term with care. / Thread ends.

Share this Scrolly Tale with your friends.

A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.

Keep scrolling