With U.S. officials openly blaming Russia for hacker attacks on state election computer systems, and the myriad possibilities for election chaos such attacks raise, it’s important to put them in proper context. I went to Harri Hursti, a globally-known election security consultant, for some answers. Hursti cut his teeth in the Finnish military fending off electronic attacks, so he has valuable perspective – particularly on a unique part of Russian culture which could explain who is really behind the attacks. He also explains the potential for psychological warfare in this incident, and why it all feels a bit familiar to his Cold War sensibilities.
Harri Hursti developed the Hursti Hack(s), in which he demonstrated how the voting results produced by the Diebold Election Systems voting machines could be altered. HBO turned the Hursti Hack into a documentary called “Hacking Democracy” which was nominated for an Emmy award for outstanding investigative journalism. Hursti is co-author of several studies on data and election security, and his consultancy. Nordic Innovation Labs, advises governments around the world on election vulnerabilities.
Between 1984-1989, Hursti worked for the UNESCO and the Finnish military in technology and cyber defense initiatives.
What do you think of the news that a member of Congress says there is “no doubt” that Russia is behind recent attacks on state election systems: (http://www.reuters.com/article/us-usa-election-cyber-idUSKCN1220SL)?
The article makes several dangerous assumptions about the security of elections and election systems. Representative Adam Schiff said he doubted (Russians) could falsify a vote tally in a way that effects the election outcome. He also said outdated election systems makes this unlikely, but really, it just makes it easier. The voting machines were designed at a time when security wasn’t considered, included, or part of the specifications at all.
These outdated computers are extremely slow. They don’t have the extra horsepower to do decent security on top of the job they were designed for. Basically, the voting machine is as powerful as today’s refrigerator or toaster, but some use the same components and logistics so outdated doesn’t mean it’s forgotten and obsolete, it means that it’s common and therefore a lot of people still today know how those systems work and can subvert them. “Outdated” isn’t offering any protection from an attacker, quite the opposite.
So there’s no proof of voter registration tampering?
As in voting machines, the registration machine don’t have the capability of logging an alteration, and they are trivially altered themselves. It’s meaningless to claim there’s no evidence, since the systems don’t have the capability to report when they’re altered. These are not standard parts of a database so there’s no common sense in saying “there’s some sort of feature that would do that, right?” Unless we study the system we can’t know one way or the other. This isn’t a common sense claim, this is a claim that would require a forensic investigation.
In addition, the number of vendors and different systems is low, so a skillful attacker doesn’t need to learn hundreds of systems; they only need to know a half dozen to control all of the U.S. election systems. But a skillful attacker only needs to learn one system in order to manipulate enough votes to tilt the election, even if it’s not close to tied. This means the attacker has more places to go to be strategic and instead of going to a big jurisdiction, They’ll go to 10 smaller ones with fewer resources and less (attention). But if you calculate the gap you need to fill to alter an election you go to the smaller, underfunded and less technologically savvy districts to own the state.
Also, some states too have made state-wide decisions that one system is used across the state, or jursdictions are central count only. So a statement that the US is decentralized is a false statement. It’s easy to understand why people think that, but from an attacker’s point of view, the threat model, you could not ask for an easier target. And the diversity between small jurisdictions is limited. An attacker can choose the jurisdictions based on the systems they are best skilled to attack.
How can the US be so sure it’s Russia?
It can’t. It is very hard to find from where a network attack is coming from. It is equally easy to make certain that investigators will find “the trail” which is pointing to the wrong direction. Therefore under the assumption that you’re dealing with a skillful attacker, any trail found is a red flag for the fact there are so many ways to make it virtually impossible to find the trail. Any conclusive looking trail “found” should be considered suspect. Unless it’s a false trail, you can only say we suspect them, and until you get to the real people to the level of the actual perpetrators true identities, you can’t make a conclusion as to “where” they come from.
Could it have been Russia?
We could use a working hypothesis, or a reasonable suspicion of Russian involvement, but until you’re down to individual people you don’t know who they are. They might even have been based in Russia, but have arrived there as tourists to carry out their attacks. There’s no way to know who the individual attackers are until they’re confronted.
Given your Cold War background, does this feel familiar?
The Cold War was all about ideology, and therefore a large concept was something that we today call hybrid warfare. In that game the actual technological attacks are equally important as the psychological influencing of the general population with misinformation and misdirection. So this is all very familiar.
Also, something we in the Western world don’t understand is how deeply patriotic Russians are. Individual Russians, and self-organized groups, are willing to go to great lengths on their own, with their own initiative, if they believe that what they do will benefit Mother Russia, and/or in hope and believe that their actions once known will be rewarded. So this kind of self-initiated actions which do resemble organized operations are commonplace. Bearing in mind that the self-organized groups can have members whose day jobs are close to the government the remaining question is, is the government aware of these groups, and if they are, are they encouraging or discouraging? Which is something we cannot know. But the fact of the matter is that Russia is self-organizing and self-providing the capability of plausible deniability which in many case can be actually true that they didn’t know.
Also, it is good to understand how high is the level of science education in Russia and the eastern bloc, when East Germany and West Germany united they had to tone down the science education in East Germany in order to match the West Germans. Science education in East Germany was way higher. The percentage of people in the general population of Russia who possess the relevant skill sets for carrying out this kind of attack is higher than we assume based on western standards. And that’s not just Russia but the whole Eastern Bloc, it was very high and is still.
Given the number of, say, smartphones and laptops used today how hard is it to fend off an attack?
In today’s world where we have “bring you device” models everywhere, we inherently assume every risk the wireless world brings to us. Our laptops and mobile phones are paired to our home networks and other wireless places we visit. It is still not understood how little security WiFi has and how easy it is with an “evil” access point to gain a connection to a target, and once you have a target you can start to work to gain access.
To mitigate this we have two possible paths. One path is to ultra-high security with all the restrictions it comes with. The alternative is to assume a breach is imminent and utilize experts to put in place an active defense mechanism which catches the breach before the attacker can use the breach to gain access to valuable information.
What would an appropriate US response be if the U.S. discovered foreign hackers in its election system?
The first action is obviously to secure your home base. Taking into account the difficulty of identifying the actual attacker, a public retaliation towards an assumed attacker may be part of the attacker’s plan and intensify the attack. Hence, public retaliation is not an effective defense. Public disclosure is important, but after the fact and after the situation has been properly handled.
Finally what is the real risk here? Could Russian hacking throw the Nov. 8 result into doubt? Could Trump supporters, should they lose, blame Russia, for example?
There’s a myriad of risks. Just to start from the simple fact that attacking the voter registration system is a highway to all crimes involving identity theft. Therefore, massive breaches of voter registration databases might lead to discouragement of people to participate in the democratic process and cause them to drop out by ceasing to be registered voters. It also poses a national security level threats, by allowing malicious actors and adversaries to gain valuable intel whether it is personal-level attacks or whether it is for hybrid warfare psyops.
It is also important to understand that data theft takes the public interest. but detect injection or insertion is far more serious. In this attack, the U.S. could be a set up for later attacks and set up false identities to be leveraged for multiple purposes in and out of the election space.
For example, a voter registration database interacts with a lot of government databases, such as criminal records. While common sense might say this kind of interaction should be a one way street, in reality the implementations quite often allow two-way interaction between the data sources. Therefore from one jurisdiction to another it should be carefully analyzed what kind of data propagation inserted voted records could lead to. Remember only US citizens can be voters, so a registered voter is assumed to be a citizen already.