On Thursday, Democrats on the House Intelligence Committee published more than 3,500 Facebook and Instagram ads linked to the Russian propaganda group Internet Research Agency, making it the largest trove of these ads the public has seen to date.
Since last fall, when Facebook disclosed it had sold political ads to Russian actors in the run-up to the 2016 election, details about the IRA’s handiwork have trickled out mainly in the form of independent research, individual exhibits presented during congressional testimony, and an indictment of IRA officials by special counsel Robert Mueller. Facebook, meanwhile, has declined to share a comprehensive list of accounts and Pages associated with the IRA on Facebook and Instagram.
With Thursday’s disclosure, the House Democrats are painting a fuller picture, which they hope will help assist in further research about the IRA’s extensive operation.
“Ultimately, by exposing these advertisements, we hope to better protect legitimate political expression and discussions and better safeguard Americans from having their information ecosystem polluted by foreign adversaries,” wrote representative Adam Schiff, ranking member of the committee, in a statement. “We will continue to work with Facebook and other tech companies to expose additional content, advertisements, and information as our investigation progresses.”
The newly revealed ads, some of which are redacted to protect the privacy of innocent Facebook users, track closely to what’s already known about the IRA’s tactics. The ads date back to early 2015 and continue through August of 2017, covering topics including LGBT rights, Black Lives Matter, immigration, Islam, veteran issues, the Second Amendment, Texas secession, and the presidential candidates themselves. The Pages implicated in this disclosure largely mirror a list published last year by the Russian media outlet RBC, which until now has been the largest list of suspected accounts published anywhere. But there are new names in the House committee’s collection, too, including an account called “Black guns matter,” which had more than 4,000 likes in November of 2016.
As was clear during the congressional hearings with tech giants last November, the IRA’s ads on Facebook and Instagram often staked out both sides of the same issue. One ad, created by the page United Muslims of America in June of 2016 and targeted at people whose interests on Facebook included Hillary Clinton and the Muslim Brotherhood, showed Clinton smiling with a woman in a hijab. It invited people to an event entitled “Support Hillary. Save American Muslims!”
Another ad, meanwhile, purchased by the page Stop A.I.—short for Stop All Invaders—showed a photoshopped President Obama in the Oval Office, with an ISIS flag behind him. “Obama was always a mere pawn in the hands of the Arabian Sheikhs,” it reads in part. “All these refugees, which we are about to take in, are soldiers with one simple goal. They are going to try to terrorize the nation.”
Some of the ads promote real-world events organized by unwitting Facebook users. One such event, promoted by the page Black Matters, promoted a night of free legal help for immigrants. Another, published by the page Fit Black, promoted free self-defense classes run by a club called Black Fist. Buzzfeed News previously reported that IRA trolls had lured American fitness trainers to lead these classes, even going so far as to pay them $320 a month.
What made these ads so deceptive is they rarely looked like traditional political ads. Often, they don’t mention a candidate or the election at all. Instead, they tear at the parts of the American social fabric that are already worn thin, stoking outrage about police brutality or the removal of Confederate statues.
Throughout Congress’s investigation into Russian meddling in the election, legislators have questioned tech leaders about how the IRA targeted the ads. Facebook has repeatedly said that the majority of ads were geographically targeted broadly to the United States, and the House’s trove backs that up. But some ads in the new collection, including a few published by the page Black Matters, specifically target cities with a history of headlines about racial unrest and police brutality, like Baltimore, Maryland, and Ferguson, Missouri. Others target by interest category. In January of 2016, a single ad for a page called Williams&Kalvin, two YouTubers the IRA hired to promote news and video about the black community, targeted users age 18 to 45 who were interested in BlackNews.com, the color black, or HuffPost Black Voices but were not Hispanic or Asian American. Facebook says it has since revisited some of its targeting categories, eliminating one-third of the categories the IRA used.
Williams&Kalvin appears to be one of the pages that tested which messages would resonate most with audiences. Within a single hour on January 14, 2016, that page posted the same ad with two different messages. The ad that read “Community about black social and racial issues! Like to subscribe!” received two clicks. The ad that read “Black Discrimination Awareness! Like to join!” got 2,592. That kind of testing is common in digital advertising, but it suggests the IRA was operating at a certain level of sophistication.
Facebook has sought to downplay the damage caused by these ads, often emphasizing that the IRA spent just $100,000 on them, compared to the nearly $40 billion in ad revenue the company earned in 2017. But the House’s disclosure shows how far even a small sum can go. One ad, purchased on June 23, 2015, by the page LGBT United, cost 99.95 Russian rubles, or the equivalent of $1.59. That ad garnered 26 clicks and 374 impressions in a single day. Looked at this way, Facebook’s argument that 50 percent of the ads cost less than $3 doesn’t sound all that compelling.
The ads also represent just a sliver of what the Russian trolls posted on Facebook and Instagram. While 3,500 sounds like a sizable number, IRA accounts published some 80,000 organic posts on Facebook and another 120,000 on Instagram, reaching roughly 146 million Americans between the two.
Since these ads ran—and the world found out about it—Facebook has announced significant changes that it says could ward off such malicious behavior in the future. In advance of this announcement, Facebook sent a bulleted list of these changes to reporters. Going forward, for instance, it’s requiring all political advertisers, including people advertising about hot-button issues such as abortion and gun rights, to provide a copy of their government-issued identification and an address to verify that they live in the United States. It will require similar verification for popular pages.
The company will also begin labeling political ads as such, and, in June, it will launch a repository of political ads that includes information on the amount spent on the ad, its reach, and the demographics of the audience it targeted. That will eliminate the kind of blind confusion that has ensued over the last year regarding the very ads the House made public Thursday. Meanwhile, Facebook will scale its content moderation team to 20,000 people by the end of this year.
The House Democrats say they plan to eventually publish the organic posts created by the Russian trolls as well. But even on its own, Thursday’s data dump could serve as a useful guide not only for researchers who are still mapping out the IRA’s tactics but also for other tech platforms—like Reddit and Tumblr—that didn’t detect Russian trolls until far later than their larger counterparts. The goal, after all, is not just to recap what happened but to prevent it from happening again.
Update: 10:10 am ET 05/10/18 This story has been updated to reflect that after publication, House Democrats revised the total number of ads from more than 3,300 to more than 3,500.