U.S. Elections Are Safer from Foreign Interference, But Gaps Remain
Four years after a big wake-up call, federal, state, and local governments are working harder and more closely to ward off threats.
If 2016 was a wake-up call about foreign interference in U.S. elections, 2020 is the acid test of the government’s efforts to better spot and ward it off. There are tangible signs of success — but also of ever-growing threats.
Within a year of the election, the Department of Homeland Security had designated election infrastructure as critical infrastructure; within two, it had created the Cybersecurity Infrastructure Security Agency, or CISA, to monitor threats. In August, CISA director Christopher Krebs called election security one of the agency’s top five priorities.
Since a lot of what happens in national elections happens on the state and local level, that means a lot of outreach to local election outfits to prepare them for foreign interference. That preparation can happen in a variety of ways, including through massive exercises, the most recent of which took place in August and brought together federal, state, local and private sector infrastructure owners and operators to role-play different election disruption events.
If states need cyber help around Election Day, the National Guard could respond, as they did in October after a ransomware attack on some government offices in Louisiana. These National Guard units are backed up by U.S. Cyber Command under a year-old program called Cyber 9-Line that helps Guard units quickly tell the command about an incident or attack, and receive unclassified data back that they can then share with local authorities and officials. The FBI, DHS, and other government agencies can share the 9-Line data too, or contribute to the exchange.
Programs like 9-Line “close the gap between when we know about adversarial activity and when we can do something about it,” Army Brig. Gen. William J. Hartman, who commands the Cyber National Mission Force, told the audience at an October CISA forum.
For the most part, actually manipulating vote totals requires physical access to voting machines, which have long been more vulnerable to manipulation than their manufacturers have admitted. (Only this year did a major voting-machine maker finally consent to working with academic penetration testers in a bid to better secure their products.) Yet while elections researchers have occasionally found, say, voting machines left unattended in polling-place hallways, it would be hard to do at scales large enough to sway an election. More worrisome are voting systems connected to the internet. In 2017, CISA officials testified to Congress that voting machines “are not connected to the internet.” Researchers subsequently found dozens of systems that were.
“There is no evidence a foreign adversary has gained access to election infrastructure. Given the size, complexity & diversity of America’s electoral system no country has the ability to change the outcome of the election,” a senior Defense Department official said on Monday.
Information related to individual voters is somewhat more at risk, however, since it’s stored on computers and networks that are accessible over the internet. In August 2016, the FBI issued an alert that Russia had breached voter registration databases in Arizona and Illinois. Four years later, Russia’s behavior doesn’t seem to have changed much. On Oct. 22, CISA officials announced that Russia had "exfiltrated data from at least two victim servers" and "there may be some risk to elections information housed on [state, local, territorial, and tribal] government networks." They didn’t say exactly what kind of information.
But CISA saw no indication the integrity of election data was compromised — that is, that Russia or anyone else has actually altered these databases, which could disenfranchise some voters. And while foreign actors may have obtained voter information, there’s no indication that they have any info that couldn’t be purchased legally (albeit, likely without the knowledge of the individual voter).
Disinformation and influence campaigns
Beyond threats to voting machines and voter databases, foreign actors have also sought more generally to inject division and uncertainty into democratic processes through disinformation, hack-and-leak campaigns, efforts to instigate or amplify civil strife or discord or simply undermine faith in the integrity of the election. That’s part of the reason officials have been doing media appearances, holding briefings, and issuing alerts to assure the public that small disruptions — say, a DDOS attack that knocks a state elections site offline by flooding it with traffic — don’t actually indicate a compromised vote.
On Monday, Sen. Mark Warner, D-Virginia, warned that influence attacks are still a threat. “Folks: this is an unusual election,” tweeted the vice chair of the Senate Intelligence Committee. “Our intelligence community has warned that the period immediately before and after Election Day is going to be uniquely volatile, and our adversaries will seek to take advantage of that. Don’t make their jobs any easier.”
The NSA / CYBERCOM has established a small group to discourage the sort of trolling the United States saw out of Russia. On Election Day in 2018 — the midterm elections — this group disrupted Internet access for Russian influence actors. In the runup to the election, they expanded it to include China, North Korea and Iran.
Larger social media companies like Facebook and Twitter have also gotten better at detecting and purging foreign actors waging influence campaigns and are making special efforts this year to directly counter misleading information, such as, for Facebook using (unspecified) tools it uses in places like Myanmar where election-related violence is a possibility. Twitter has added face-checkers and labels for some tweets that might contain misinformation. But some deceptive content can still make it onto the site and receive views in the hundreds of thousands or higher before Twitter labels it.
Moreover, federal authorities from various agencies are cooperating better to detect and counter attacks. In 2016, “You had pockets of folks; some folks would be following one stream of information; another group might be following a different stream of information,” Cynthia Kaiser, a section chief with the FBI Cyber Division, said at the October CISA summit. “Really, what we thought about after 2016 was ‘how do we sync up all of these kind of disparate groups that want to do the right thing…to make sure we can counter these threats holistically?’”
Kaiser said the FBI created a foreign influence task force “to really look at the influence threats overall.” It includes members of the bureau’s counterintelligence, counter-terrorism, and criminal-investigative communities. Her own cyber team has designated election leads both for operations and intelligence, she said.
Unlike in 2016, there is much better and more active coordination with other government agencies, such as CISA and the NSA. Kaiser said she feels “heartened” by the intelligence that they get from victims of hacks and intrusions. “The majority of the information that we’ve been able to get out is because somebody called us because they saw something suspicious and we investigated.” That, too, is a change from 2016.
Overall, that coordination has lead to much faster attribution and action against foreign interference than 2016. Consider the Wikileaks attack, which private cybersecurity company Crowdstrike attributed to Russia as early as June of that year but which the government did not publicly attribute to Russia until October.
In 2020, by contrast, the government was very quickly able to point out an Iranian influence operation aimed at Florida voters only hours after it occurred, even if the messaging around the incident was somewhat confused.
Foreign attackers do still have some soft targets to aim for in order to steal data that could play a role in influence operations. Michael Kaiser, the president and CEO of the Defending Digital Campaigns, told the CISA summit audience that third parties that are tied to campaigns, such as contractors, etc. are an issue of some concern. “If they’re a campaign of seven people running a [U.S.] House campaign, you know, it all falls on the campaign manager mostly. Maybe there’s someone who knows a bit about IT that gets assigned some of this...To get their bandwidth and their time to engage in anything related to cybersecurity, even if they have a heart for it, which many of them do, is still difficult.”
Katie Bo Williams contributed to this report.