Investigators want to know more about how Russia used Facebook, Twitter, and Google to target ads and spread propaganda. New details raise questions about whether the tech giants should change their ad infrastructure ahead of the 2018 elections. One expert said any major hack could have given Russia the voter data it needed to target the ads.
Facebook ads purchased by Russian operatives during the 2016 election targeted certain demographic groups and geographic locations, raising new questions about how foreign actors were able to home in on key voters in swing states by exploiting the same advertising tools used by American political operatives last year.
The Washington Post reported Monday that Russian operatives had used Facebook’s custom-audiences tool to track and target users who had visited fake pages set up by the Russians to peddle disinformation and propaganda.
They also used other advertising tools to target users by demographics and geography, according to The Post and CNN. The ads targeted users in Michigan and Wisconsin – two states Trump won by approximately 10,000 votes and 22,000 votes, respectively.
Facebook’s chief security officer, Alex Stamos, said in September that “about one-quarter” of the ads purchased by Russia-linked accounts “were geographically targeted.” The company gave lawmakers the ads late last month after providing them to special counsel Robert Mueller.
The social media giant is not the only one under scrutiny by investigators probing Russia’s election interference. Google and Twitter have also been invited to appear publicly before the House and Senate intelligence committees to explain how Russia used their platforms to target ads and spread propaganda.
Google is examining how Russia may have exploited services like Google News, Gmail, and YouTube.
Like Facebook, Google’s ad platform has self-service options, and Russian actors could have tried to manipulate the search engine’s algorithm to place bogus stories at the top of search results. Gmail was similarly vulnerable. Hillary Clinton’s campaign chairman, John Podesta, had his Gmail account breached by Russian hackers during the election.
Russian news network RT, meanwhile, has 2.5 million YouTube subscribers and was sold Google’s premium ad inventory, known as YouTube Preferred, until Google pulled the network from the program last week, according to Bloomberg.
Twitter may have been used even more extensively than Facebook, experts have found, because of how cheap and easy it is to create automated accounts that have been programmed to retweet certain keywords, hashtags, and topics en masse.
The bots homed in on swing states, too. A study conducted by Oxford’s Computational Propaganda Project concluded that the automated accounts also helped spread junk news and disinformation more heavily into highly contested districts in the days leading up to the election.
But lawmakers say top Twitter representatives interviewed by the Senate Intelligence Committee were less than forthcoming about the Russia-linked accounts they had uncovered. The committee’s vice chairman, Democratic Sen. Mark Warner, told reporters later that Twitter’s testimony was “deeply disappointing” and “frankly, inadequate.”
- REUTERS/Aaron P. Bernstein
‘These methods are sophisticated, but Facebook makes it easy’
The committees probing Russia’s use of social media during the campaign also want to know whether they had any help in micro-targeting their political ads, which aimed to sow division and undermine Democratic nominee Hillary Clinton.
But experts say the companies’ ad infrastructure made it easy even for foreigners to determine whom to target, and with what kind of content. ProPublica, for example, outlined last month how easy it was to use Facebook’s self-service ad-buying platform to target specific users with racist and anti-Semitic content. BuzzFeed discovered it was just as simple with Google.
The Russians’ social media efforts were so straightforward that they could be compared to a legitimate campaign data operation run by Americans.
The Trump campaign’s digital operations chief Brad Parscale, for example, poured virtually his entire budget into Facebook, bombarding likely supporters with ads and “dark posts” that were visible to certain users but not others.
“These methods are sophisticated, but Facebook makes it easy to use them,” said Zeynep Tufekci, an associate professor at the University of North Carolina at Chapel Hill who specializes in technology and information.
Tufekci said the targeting would have been “more effective with a better understanding of US politics.” But, she added, “it’s crucial to note that many American political actors and others also used these tools – sometimes to target misinformation and fake news.”
“The targeting tools make it easy to figure out who who liked the message or clicked, and go from there,” Tufekci said.
‘Too amenable to spreading misinformation’
- Thomson Reuters
Short of special counsel Robert Mueller or the congressional intelligence committees discovering coordination between the campaign and Russia’s digital operations, then, the problem now lies at the feet of Silicon Valley.
By making it so easy to purchase, target, and hide the source of political ads, the tech giants may have unwittingly facilitated a sophisticated influence operation capable of influencing the vote – especially in swing states like Michigan, New Hampshire, Wisconsin, and Pennsylvania.
Hackers linked to Russian military intelligence also set up Facebook accounts called “Guccifer 2.0” and “DCLeaks” that they then used to disseminate the emails stolen from the Democratic National Committee, according to the Washington Post. Facebook alerted the FBI to the accounts before shutting them down.
“Any kind of hack,” moreover, could have supplied the Russians with the voter data they needed to feed into Facebook’s ad targeting tools, said propaganda and misinformation expert Jonathan Albright.
Voter databases infiltrated by Russian hackers last year would have been another treasure trove of data for anyone looking to target political messaging and advertising. The National Security Agency concluded earlier this year that “Russian intelligence obtained and maintained access to elements of multiple US state or local electoral boards” in the days and weeks leading up to the election.
That kind of voter data is highly valuable, experts say. But Facebook has denied that any of the ads used the type of Custom Audiences targeting that would require personal information like email addresses. The voter data wouldn’t have been essential for the ad targeting tools, anyway, “based on how Custom Audiences can already geo-target, segment, and track audiences’ behaviors and media consumption across the internet,” Albright said.
Facebook, for its part, said it plans to strengthen its review process for political ads and will begin disclosing which pages purchase them moving forward. But the company acknowledged last month that, because many of the ad purchased could be made electronically with no human interaction, it still doesn’t know the full extent of the ads purchased by Russia-linked accounts.
Ultimately, Tufekci said, “the Russia scandal exposes the surveillance and ad infrastructure that lies at the heart of Silicon Valley business model that is unfortunately too amenable to spreading misinformation, fake news, and deceptive propaganda to wide audiences – easily and cheap.”