An extensive disinformation campaign targeting Democracy 2020 candidates is already underway on social media, indicating that foreign state actors are at least driving part of the business.
The main targets appear to be Senator Kamala Harris (D-Calif.), Senator Elizabeth Warren (D-Mass.), Former Representative Beto O. Rourke (D-Texas) and Bernie Sanders (I-Vt. ), four of the largest candidates announced or potential to the presidency.
History continues below
A review by POLITICO of recent data extracted from Twitter and other platforms, as well as interviews with scientists and digital campaign strategists, suggests that the goal of the coordinated dam seems to undermine nascent candidacies by spreading memes, hashtags and misinformation. , and distortions of their positions. But the conflicting nature of many positions has also allusion to a broader effort to sow discord and chaos in the Democratic presidential primary.
Cyber-propaganda – which is often about the most raw and sensitive topics in public discourse – is pushed through a variety of platforms and with a more insidious approach than in the 2016 presidential election, when online attacks intended to polarize and mislead voters have surfaced for the first time on a large scale.
Recent publications that have been widely disseminated include racially inflammatory memes and messages involving Harris, O'Rourke and Warren. In the case of Warren, a false story surfaced, alleging that a blackface doll had appeared on a kitchen cabinet in the background of the Senator's Eve's Instagram video stream.
All activity is not organized. Much seems to be organic, reflecting the politically polarizing nature of some candidates. But there are clear signs of a coordinated effort of indeterminate size sharing the same characteristics as the computer propaganda attacks launched by trolls online at The Russian Internet research agency in the 2016 presidential election, which special advocate Robert Mueller has accused of aiming to undermine the political process and elevate Donald Trump.
"It seems that the 2020 presidential primary will be the next battleground to divide and confuse Americans," said Brett Horvath, one of the founders of Guardians.ai, a technology company that works with a consortium of scientists, academics and computer technologists. to disrupt cyberattacks and protect pro-democracy groups from the information war. "With regard to the information war in the 2020 cycle, we are not on the eve, we are already in the third inning."
An analysis conducted for POLITICO by Guardians.ai showed that a relatively small group of accounts – and a larger group of accounts that amplifies them – has conducted a disproportionate amount of Twitter conversations on the four candidates in the last 30 days.
Using proprietary tools that measured the debate around candidates in the Democratic realm, Guardians.ai identified a cohort of about 200 accounts – including both involuntary real accounts and other "suspicious" accounts and automated coordinating to broadcast their messages – rejected negatively. or extreme themes designed to harm the candidates.
This is the same group of accounts the company identified for the first time last year in a study a large-scale influence campaign in the 2018 elections.
Since the beginning of the year, these accounts have begun to focus specifically on Harris, O'Rourke, Sanders and Warren, and have been magnified by an even larger group of accounts. Over a recent 30-day period, between 2% and 15% of all Twitter endorsements of the four candidates came from one way or another from this group of accounts, according to Guardians.ai's findings. During this period, the four candidates collectively had 6.8 million mentions on Twitter.
"We can conclusively assert that a large group of suspicious accounts that participated in one of the most important influential transactions of the 2018 cycle is now engaged in sustained and continuous activity for the cycle. 2020, "said Horvath.
Amarnath Gupta, a researcher at the San Diego Supercomputer Center at the University of California, San Diego, monitors social media activity and recently witnessed an upsurge in Twitter activity targeting three candidates – O'Rourke, Harris and Warren.
This increased activity includes an increase in the number of tweets, the speed with which they are posted, and the appearance of a "clustering behavior" related to the three candidates.
"I can say that very quickly, a lot of information is biased with regard to the analysis of feelings," said Gupta, a Guardians.ai partner in a study conducted in 2018.
According to Guardians.ai's analysis, Harris would have been the most comprehensive Twitter activity among the 2020 candidates reviewed, with over 2.5 million mentions over a 30-day period.
She was also among the most targeted. A widespread tweet has used racist and sexist stereotypes to try to sensationalize the relationship between Harris and former San Francisco Mayor Willie Brown.. This tweet – and the retweets and mentions related to it – generated 8.6 million "potential impressions" online, according to Guardians.ai, a calculation of the upper limit of the number of people who could have see according to the accounts that the cluster follows, which follows the accounts within the group and which is engaged in the tweet.
Another racist tweet was directed to O'Rourke. The Twitter profile of the user who created it indicates that the account was created in May 2018, but it has only written one tweet since – in January when the account announced that he had breaking news about the fact that a former Texas congressman had left a message using racist language. on an answering machine in the 1990s. This tweet generated 1.3 million potential impressions on the platform, according to Guardians.ai.
A separate study by Guardians.ai, which examined the interest of the group of 200 accounts in electoral fraud and false and / or deceitful accounts of electoral integrity, published just before the mid-term and co-election elections. Written by Horvath, Zach Verdin and Alicia Serrani, reported that accounts generated or were mentioned in more than 140 million tweets compared to the previous year.
This group of accounts was the driving force behind an effort to aggressively advance conspiracy theories in mid-2018, ranging from misinformation about electoral fraud to stories involving a caravan arriving in the United States and even the defense of violence.
Horvath says that the activity around the cluster represents an evolution of misinformation and amplification tactics that began in the middle to the end of 2018. The initial phase that began in 2016 has been marked by the creation of thousands of more easily detectable accounts as robots or as a coordinated activity.
The new business, however, is focused on a refined group of key accounts – the same accounts that were revealed in the group's 2018 election fraud study. Some of the accounts are thought to be very sophisticated synthetic accounts run by people trying to influence conversations, while others are coordinated in one way or another by actors who have identified real people already tweeting the desired message.
Tens of thousands of other accounts then work together to amplify the core group through mentions and retweets to drive what seems, on the surface, to be an organic virality.
Digital business operators, political campaigns and other social media monitoring groups also report a recent increase in false stories or negative memes against 2020 candidates.
A recent analysis Storyful detected spikes in misinformation activity on social media platforms and online comment tables a few days after each of the 2020 candidates ran for president, starting with Warren's announcement on 31 December. December.
Storyful found that secondary information websites and ancillary social media platforms played an important role in spreading the anti-Warren sentiment in the days that followed the announcement of his candidacy on December 31st. Using various keyword searches for Warren's mention, the company has reported evidence of "Spam or bot" as an activity on Facebook and Twitter from the best posters.
Kelly Jones, a researcher at Storyful who detected suspicious activity during the three days following the announcement of the Harris election campaign, Warren, Rep. Tulsi Gabbard (D-Hawaii) and Senator Cory Booker (DN.J.) said they saw a concerted push on separate bulletin boards to create false or derogatory stories.
Among the marginal platforms identified as Storyful, there were 4Chan and 8Chan, where messages prompting commentators to quietly ravage Warren on social media or in the comments section under News.
"Point out that she was Republican, but she changed sides and became a spy for them now. Use this quote out of context: "I was a Republican because I thought it was those who best supported the markets," wrote a poster on 4Chan's bulletin board.
"We see a lot of this rhetoric for almost every candidate who comes forward," Jones said. "There is a call for action on these peripheral sites. The terrain will be so crowded that they will say "OK:" Split operation left. "
A Harris campaign official said they suspected of bad actors pushing for misinformation and that false stories about the California Democrat were trying to divide African Americans or to get the media to pay attention disproportionate to critics designed to create divisions among the core Democratic electorate.
Researchers and others interviewed for this story say they can not conclusively identify the actors behind the coordinated activity. We do not know if they are dishonest hackers, political activists or, as some claim, foreign players, such as Russia, because they bear the mark of previous foreign attacks. One of the goals of the activity, they say, is to divide the left by making the Democratic presidential primary as chaotic and toxic as possible.
Teddy Goff, who served as Obama for the US digital director, generally described ongoing efforts as "a mishmash." It's a bit of an unholy alliance. "
"There are supporters and funders of the state in this area. Russia. North Korea is believed to be one, and Iran is another, "he said. "In some cases, it seems coordinated, but whether it is coordinated or not, there are clearly actors who are trying to influence the primary by exacerbating the divisions within the party, by presenting more moderate candidates. as unpleasant for progressives and more progressive candidates as unpleasant for more traditional Dems. "
A senior Sanders campaigner expressed "serious concern" about the impact of misinformation on social media, describing the phenomenon as "a cyber-political war with a clear impact on the democratic process." the activity that he sees already as involving both foreign and domestic actors.
Twitter and Facebook, which owns Instagram, have said they have taken significant steps since 2016 to identify and block foreign players and those who violate the rules of the platform.
Although Twitter does not specifically answer questions about the findings of Guardians.ai, last year the company reported Challenge millions of suspicious accounts every month, including those with "spammed and automated behavior". After trying to authenticate the accounts by email or phone, Twitter suspended 75% of the accounts put in competition from January to June 2018.
In January 2019, Twitter released a accounting efforts to combat foreign interference on political conversations taking place on the platform. Prior efforts included release datasets potential foreign news operations appeared on Twitter, including 3,841 accounts affiliated with the IRA, originating in Russia, and 770 others, potentially from Iran.
"Our investigations are global and ongoing, but the datasets we've recently released are the ones we're able to reliably assign and disclose," a Twitter spokesperson said in a statement. at POLITICO. "We will share more information, if and when it is available."
Facebook says that 30,000 people work in the security field and that it blocks and removes more and more fake accounts. the the company also says he brought unprecedented transparency to political advertising on his platform.
At this early stage, the campaigns themselves seem ill-equipped to handle the attack online. Their digital operations are geared towards fundraising and organizing while their social networks are designed to communicate positive information and messages. Although some have resorted to surveillance practices, defense measures usually take a back seat, especially since the source and scope of the attacks are still unknown.
A high-profile player for a leading candidate in 2020 highlighted the enormous challenges facing campaigns, even those with the most sophisticated digital teams. The problem already seems much more important than the resources available to a candidate at the moment, said the official.
Alex Kellner, managing director of Bully Pulpit Interactive, The number one digital firm for Hillary Clinton's 2016 campaign warns that campaigns that lack a serious infrastructure to fight misinformation and dictate their own online messaging will be the most vulnerable to attack in 2020.
"I think it will be an integral part of any successful campaign: monitor it and work with the platforms to stop the misconduct," Kellner said.
Kellner said that even though platforms such as Twitter and Facebook have stepped up their internal efforts to find bad actors, the flood of false information and misinformation attacks against the 2020 candidates is already important.
"All the infrastructures we saw in 2016 and 2018 are already fully operational. And by 2020, the situation will only get worse, "said Kellner, pointing to the negative memes that were attacking Warren about his claims regarding the Native American heritage and the memes surrounding Harris's relationship with Brown.
The proliferation of false information, the rapidly evolving techniques of malicious actors and an under-prepared number of Democratic candidates could create an extremely volatile election season.
"The moderates, the centrists and the Democratic candidates still do not understand what happened in 2016 and they did not realize, like Hillary Clinton, that" we are in the middle of the day. she was not only leading a presidential campaign, she was involved in a global war of information, "Horvath said. . "Democratic candidates and presidential candidates from the center and from the right who do not understand that the campaign will not be difficult, they will let their campaign be an unintentional booster of the attempts of someone else to further divide Americans. "