Ever since the 2016 election it’s become clearer that ongoing Russian disinformation tactics on social media seek to exploit existing fault lines in American society. Last week, even as the Justice Department announced new charges against a Russian operative running a well-financed conspiracy to interfere with the midterm elections, two new reports exposed how Facebook and Twitter serve as platforms to target black Americans specifically, stoking domestic racism and discord.

First came Twitter’s release of a dataset of nearly 10 million tweets from 3,800 accounts that the company says were operated by the Kremlin-backed Internet Research Agency (IRA). This trove of data caused researchers at the University of Washington to reconsider the role Russian propagandists may have played in dynamics those researchers previously observed on the network. With the list of newly identified Russian accounts in hand, researchers including Kate Starbird, Ahmer Arif and Leo Stewart combed over datasets assembled to look at activity involving Black Lives Matter, and for the first time were able to distinguish the Russian effort clearly.

Starbird and her colleagues study how social media discourse evolves. Armed with the new dataset of IRA accounts, they reconsidered the results from a recent paper that dealt with the framing of Twitter arguments around racial justice and social movements. Generally, the paper explored how social media networks allow for framing and gatekeeping around ideas and issues, allowing groups to define and collectively produce and amplify ideas related to the discussion.

The UW researchers looked at tens of millions of tweets in their analyses. Perhaps it is unsurprising to see that clusters emerge among left-leaning and right-leaning groups in discussions involving hashtags such as #BlackLivesMatter. But when the researchers went back to their findings with the identified IRA accounts, they could see the impact of the Russian effort on the discourse. More vividly, they could see the IRA accounts clearly visualized in their network analysis.

The IRA accounts successfully impersonated activists on both sides of the issue, creating a network of relationships with real Americans. Starbird and her colleagues see two main takeaways from their analysis:

First, these information operations are targeting us within our online communities — the places we go to have our voices heard, to make social connections, to organize political action. They are infiltrating these communities by acting like other members of the community, developing trust, gathering audiences. Second, these operations begin to take advantage of that trust for different goals, to shape those communities towards the strategic goals of the operators (in this case, the Russian government).

In the resulting paper, “Acting the Part: Examining Information Operations Within #BlackLivesMatter Discourse,” the researchers found four systematic patterns of forged profiles operated by the IRA. The first two included “‘the proud African American’ as a political identity, on the one hand, and the articulation of ‘the proud White Conservative.’” The third and fourth patterns presented organizational accounts that pretended to be grassroots political and media groups representing both sides.

The two sides generated believable content largely by referring to real events and facts. The IRA used accounts from both sides of the political spectrum to attack the veracity of the “mainstream” media, seeking to undermine the role of the American news media as information gatekeeper.

These findings are consistent with  another recent report that reviews Facebook advertisements placed by the IRA. Last week, the social justice watchdog group Stop Online Violence Against Women (SOVAW) published a preliminary report examining a dataset of 3,500 IRA Facebook ads released by Congress last year. In the report, “How The Facebook Ads That Targeted Voters Centered on Black American Culture: Voter Suppression Was the End Game,” researchers led by SOVAW founder Shireen Mitchell observed that the IRA advertisements chiefly focused on black identity issues, and assess that these efforts led to voter suppression. Ads invoked historical civil rights figures such as Malcolm X and Martin Luther King Jr., alongside issues such as police brutality, slavery and discrimination.

SOVAW created a linked graph of the 3,500 Russian IRA ads. Each node in the graph represents a topic or issue. The color of the nodes indicates the volume of impressions, including the number of ads bought. The number of impressions are color-coded, low (red) to mid-level (yellow) to high (blue) The clicks are visualized by size of the node- the larger nodes received more clicks.

The SOVAW report shows the Russian focus on black identity and civil rights issues was complementary to the Trump campaign’s voter suppression effort in 2016. It notes that Brad Parscale, the Trump campaign digital director, has openly claimed Donald Trump’s bid for the White House included Facebook “dark posts” delivered to suppress black voter turnout. And, it argues that targeting black women specifically was a dreadfully expedient target for the Russians and the Trump campaign: since it is relatively easy to attract online harassment of black women (as evidenced by campaigns such as #StopBlackGirls) and because black women tend to turn out in high numbers to vote, they are an obvious constituency to suppress.

“This report from SOVAW reveals the extent to which Black American culture was specifically targeted, with the explicit goal of suppressing voter turnout,” said Whitney Phillips, Assistant Professor of Communication, Culture, and Digital Technologies in Rhetorical Studies at Syracuse University, in a public statement about the report. “It also contextualizes these efforts alongside racist media manipulation and disinformation campaigns spanning a half-decade. The report emphasizes the persistent threat these efforts pose to black voter turnout in the upcoming 2018 midterms and beyond.”

Russian exploitation of America’s race divisions is not new, but the efficiency and scale of social media platforms to target vulnerable groups has changed the nature of the threat. These reports make clear that Facebook’s and Twitter’s repetitive failures to address state-sponsored disinformation on their platforms have created a profound new threat to the advance of social and racial justice in the United States. Civil rights groups –already beleaguered by the fight against an onslaught of regressive policies of domestic politicians — must also now fight well-funded state actors who seek to use their interests as leverage in a broader campaign to destabilize the West. It is crucial that lawmakers take this threat seriously, and that the social media companies take urgent action to protect vulnerable groups. Hopefully the Russian incursion into both left-leaning and right-leaning groups will encourage more political representatives to address these threats to public discourse. Hopefully the companies will be motivated at least by the growing consumer distrust of their platforms to address these problems and invite outside experts to help in that cause.