Did Russians Hack US Voting Machines? Nobody Knows.
US authorities say they have no evidence — but experts say there’s been no investigation. The truth is that no one knows, and claiming it didn’t happen is just as misleading as claiming it did.
This morning on Twitter, journalist Adam H. Johnson took issue with a New York Times reference to Russia “hacking” the 2016 election. Johnson called the phrase misleading, claiming it “gives readers the impression voting machines were hacked.” He then cited an Economist/YouGov poll that showed more than half of Democrats believe Russians tampered with voting machines, the implication being that this was incorrect information.
It is not, as far as anyone knows.
I disagreed with Johnson in brief exchange, reminding him that US intelligence sources say Russian hackers accessed desktop computers at Boards of Election in 39 US states. Johnson responded that this is merely an allegation (which is true — but then so is everything we “know” about Russian tampering in the 2016 election) and that Boards of Election do not impact vote totals. This latter is factually incorrect — more on that shortly.
When I pushed, Johnson repeated the oft-cited refrain that US intelligence sources “have no evidence” Russians tampered with voting totals. This is a true fact, but it is utterly meaningless, because computer science experts agree that no investigation has taken place that might have turned up such evidence.
Johnson is a quality journalist who I hold in high regard; my intent is not to pick on him personally. My complaint is that journalists in the US, as a community, have grossly misrepresented our understanding of Russian influence on the 2016 election. The truth is, no one knows whether hackers, Russian or otherwise, tampered with voting machines.
Prior to the 2016 election, computer scientists raised alarms about the vulnerability of US voting machines to hacking. Popular understanding of computer science is lacking, unfortunately, so many journalists and politicians incorrectly believe that a voting machine is safe when it is not, itself, connected to the Internet.
In fact, as computer science experts like Alex Halderman and Andrew Appel tried to warn us, voting machine hacks like that employed by Russia in Ukraine’s 2014 election do not work by targeting individual voting machines. Instead, they target the web-connected desktop computers from which those voting machines are updated.
It’s true that most voting machines are not connected to the Internet. In order for the machines to function, officials must populate them with basic information like offices and candidate names. This is usually done via portable USB drive, or some equivalent, often right before the election.
Those USB thumb drives are updated from centralized desktop computers, almost always at the offices of local Boards of Elections. To hack voting machines, one need access those computers and upload a worm — a type of computer virus, essentially. The worm spreads to the USB drive (or drives) by itself, and from there to dozens or hundreds of individual voting machines. Once on those machines, the worm will alter votes to achieve the intended result, and then delete itself and erase all evidence that it ever existed.
Again, this is not a theoretical approach — this is exactly what Russia did in Ukraine in 2014. To do the same in the US (which uses many of the same voting machine models) hackers would need to access computers at Boards of Election — exactly what the Intelligence Community says happened in 39 US states.
In Ukraine, officials were concerned about the possibility of a Russian hack, and voting machines were closely watched. Technicians caught the worm in the act of changing votes, before it could delete itself, and averted a crisis. No such watchdogs exist in the United States, where politicians and media [none of whom are experts in voting machine technology] insisted before the election that the US was immune to any possibility of hacking.
If, therefore, hackers tampered with voting totals in the United States in 2016 using the same type of worm used in Ukraine in 2014, we would see no evidence after the fact. The only way to find evidence of such a hack would be a manual audit of paper ballots against electronic voting records — and no such audit has been conducted anywhere in the US. [The reviews demanded by Jill Stein in Pennsylvania, Michigan, and Wisconsin (only one of which was completed) were recounts, not audits, and would not have produced such evidence.]
Well, there might be one other bit of circumstantial evidence. Statisticians at Vote Sleuth, and other organizations have pointed to irregularities in vote totals, particularly in swing states, that they claim indicate tampering to switch electronic votes from Hillary Clinton to Donald Trump. The math behind this is too complex for me to explain — follow the link if you’re interested.
In summary, we know (at least according to US intelligence sources) that Russians favored Donald Trump. We know that they hacked the Ukraine election in 2014, using a technique that would work in the US. We know they hacked into Boards of Elections in 39 states, a critical step in employing such a worm. We know that government sources have “no evidence” of a worm, but we also know that it would have deleted itself and left no evidence, except through vote audit — which hasn’t happened.
In other words, we know bank robbers broke into our vault. We know they’ve stolen from other vaults, and we know they wanted our money. Our bank manager says he has no evidence they stole anything— but he hasn’t actually counted the money.
To be fair, other statisticians question the methodology and conclusions applied by those claiming evidence of tampering. My point here is not to tell you who is correct, but to remind you that no one actually knows, because the proper investigation has not been done.
Yet journalists, including but certainly not limited to Adam H. Johnson, mislead the public by acting as though voting machines were definitely not hacked. Why would that be?
Indulge my speculation for a moment. It’s easy to understand why Republicans would reject any notion that vote totals were changed — it would clearly invalidate the victory for their President, and possibly many other GOP elected officials. Trump, for his part, insists no Russian hacks occurred at all, and refers to the narrative as an “excuse” on the part of Democrats.
But therein might lie a motive for some journalists. Johnson might disagree with me here, but I think it’s safe to say his personal politics align more with “The Left,” a political persuasion currently devoted to the notion that the Democrats cannot win with their centrist appeals and must embrace socialism in the Bernie Sanders mold, advocating free higher education and Medicaid for all.
That thesis, and the political momentum it grants, suffers a blow if it’s revealed that Hillary Clinton rightly won the 2016 election, by both popular and electoral vote. Those pushing for a true American Left movement, therefore, might be eager to advance a narrative that hacking played only a minor role — or even to equivocate Russian meddling in the Presidential Election with the DNC’s Hillary-favoritism in the Democratic Primary.
I do note, for example, that Johnson reminds me Russian hacking of Boards of Election is “an allegation,” and yet cites with confidence claims from the same intelligence sources that they have no evidence of voting machine hacks.
My disagreement here is not even political. I am myself on the left, and very much in favor of the socialist policies espoused by Sanders and the DSA. But I’m more dedicated to knowing the truth of our systems, and the circumstances behind Trump’s victory — and especially identifying and rectifying any vulnerabilities before the next election.
It might seem politically advantageous to believe that Trump defeated Hillary because she played too close to the center. But even if such a narrative helped a strong leftist candidate win the Democratic Primary in 2020, why wouldn’t that person fall victim to the same sort of electronic chicanery? Any non-Republican candidate must already contend with voter suppression efforts, gerrymandering, and felon disenfranchisement — shouldn’t we want to know whether literal vote tampering handed Trump his victory?
I am not asking anyone, in the media or elsewhere, to advance a narrative that hackers (Russian or otherwise) changed vote totals in 2016. That would be just as misleading as saying they didn’t. What I would ask is that the narrative hew as close to the truth as possible: Nobody knows whether vote totals were changed. Anything else is just wrong.