/ #TWITTERFILES 21 How to find Russians anywhere
Pt.1 – OSPREY PROJECT –
After the 2016 election, the Senate Intel Committee asked Twitter to identify Russian Internet Research Agency accounts. But both Twitter and third-party investigators struggled to find Russian agents. Image 2/ At first, Twitter was only able to produce 22 plus 179 accounts “linked” to the IRA. Democrat Mark Warner called the numbers “inadequate across the board.” Twitter went back to the drawing board on an analytics project called “Osprey.” Image Image Image 3/ In “Project Osprey”, Twitter counted 2 types of Russians.
“A priori” Russians, identified as Russian by outside researchers such as QIntel
“Inferred” Russians (also known as is_russian): identified by an algorithm tracking “signals” such as Cyrillic text or a Russian IP address Image Image 4/ This type of analysis, based on “markers” such as the type of email carrier or retweet history, can quickly become a Rorschach test, where you see what you want to see. “If you just look for that marker, everything will look Russian,” is how one industry analyst puts it. 5/ Even Twitter has figured it out. Noting that Green Party candidate @DrJillStein was “captured by is_russian” – the list of “inferred Russians” – Twitter analysts commented on the “overbroad nature of is_russian”. Image 6/ “Incandescent,” says Stein. “Just more than Joe McCarthy would be proud of.” Image 7/ Another account considered “is_russian” was @wikileaks. Showing attribution, @Stella_Assange said: “Wikileaks doesn’t fit the definition,” noting that anyone using surveillance-resistant search tools like the Tor browser can also be considered “Russian.” Image 8/ “Using TOR would randomly result in false positives,” said Stella Assange. “TOR is an essential tool for the civil liberties and privacy communities.” Image Image Image Image 9/ In the same Osprey document, Twitter notes that the hashtag #WarAgainstDemocrats (which the NYT published 1,700 times on Election Day) only got a microscopic 6,953 impressions—one hydrogen atom in vast ocean of Twitter tweets. Image Image Image 10/ After Twitter’s first attempts to identify Russian accounts resulted in such a low number, they used different methodologies, counting an ever-increasing number of “Russians”, from 22 to 201 to 2700 to 2752 to 3124 (on Osprey) and finally, 3814. Image Image Image Image 11/ Although Twitter began to openly refer to “IRA-linked” accounts, the company had no secure way to make such identifications. They were only “potentially connected”. Image 12/ As policy chief Colin Crowell explained in late 2017, “we are citing accounts as IRAs based on third-party assessments” as “we have no realistic way of knowing based on Twitter”, other than “educated guesses”. Image 13/ Twitter knew that if they couldn’t identify IRA accounts with certainty, neither could outside investigators who didn’t have internal Twitter data.
PART 2 – Russian troll hunters
Image 14/ If you look at Racket.News today, you’ll see that MSNBC alone made hundreds of false claims of Russian interference, citing the Hamilton 68 dashboard, which past #TwitterFiles revealed to follow Americans, not the “Russian bots and trolls” as it is claimed. 15/ Outside of Hamilton 68, which Twitter kept quiet about despite being known internally to be “bullshit,” Twitter analysts may have been more dismissive of Clemson’s Media Forensics Hub, one of the main drivers of the cyber scare stories of Russian subversion in print and television. . Image Image 16/ Clemson “Troll Hunter” professors Darren Linvill and Patrick Warren appeared in one media story after another (sometimes pictured with the Hamilton 68 scoreboards in the background ), even warning that “uplifting tweets” could lead you into the dark clutches of a Russian troll. Image Image Image Image 17/ When asked how they knew an account was Russian. Warren confessed that he can’t be “100 percent sure,” but Livill said he was “certain.” He, however, would not say how. “Transparency is simply not possible,” he said. Image Image Image 18/ Clemson’s most “true” troll hunter seemed to jump the gun in 2020 when he suggested #BloombergIsRacist might be a “Russian hashtag” because it started that morning from the account @drkwarlord than “alive”.[s] somewhere in Asia” because he posts when Americans should be sleeping. Image 19/ This was fake – it started earlier that morning, and not by @drkwarlord, who is an American who lives in the USA, not Asia. When I asked why her sleep schedule didn’t seem like it, she laughed and said, “I’m a nurse at a hospital in Indiana. In 2020 I worked the night shift.” Image Image Image Image 20/ This kind of confusion seems to be what Twitter’s Aaron Rodericks was referring to when he pointed out Clemson professors’ “constant tendency to find Russians/foreign interference every time that they’re looking for something.” Image 21/ Another example occurred with the #DCBlackout hoax. While Twitter exchanged emails with the FBI, describing the campaign as “a small-scale domestic trolling effort… not a bot or a significant foreign angle,” Linvill described the campaign as having used “a classic Russian move.” Image Image 22/ Following Clemson’s subsequent report of “inauthentic activity” during the Vice Presidential debates , Rodericks denounced “their usual claims of foreign interference as baseless,” writing that they had to “scratch the bottom of the barrel to find foreign interference.” Image image 23/ But to the Clemson researchers’ credit, they discovered that Twitter was also “brushing the barrel,” having linked several Americans to a list of “Russian bots” given to Congress. So it’s clear that Twitter’s false positives didn’t stop with their project, Osprey. Image 24/ The Americans found by Linvill & Warren had been unfairly censored, suspended from Twitter without warning or explanation. Thrown into a haystack of suspected Russians, they “lost access to the Twitter accounts they used to maintain social and professional connections,” Wired wrote. Image 25/ After years of Red Scare stories (citing the Clemson duo, Ham68 and others) fueling such horrific censorship, Linvill leveled at PBS in 2022: “You know, the Russian trolls they are not as common as people think they are.” 26/ PART 3 – HOW TO FIND MISINFORMATION IN ALL THE
To be continued…
#TwitterFiles21 was co-authored with @mtaibbi and help from other #TwitterFiles researchers @mentions 27/ Clemson faculty comments on Racket News