The Seven-Second Democracy: Why India Must Go Back to the Ballot Paper

You press a button. A light blinks. Behind a small glass window, a paper slip appears for exactly seven seconds — showing you, supposedly, who you just voted for. Then it drops into a sealed box you will never open, inside a machine you will never be allowed to inspect, running code you are legally forbidden from examining.

Seven seconds. That's the entire window of transparency Indian democracy gives its own citizens.

And we're told to call this progress.

We're not the outliers here. We're the holdouts.

Here's what nobody tells you at election time: most of the rich, technologically advanced democracies that could easily afford electronic voting machines looked at them carefully and walked away. Not because they're technophobic. Because they ran the experiment and didn't like what they found.

Germany went first, and Germany went hardest. In March 2009, the Federal Constitutional Court — not a fringe activist group, the highest court in the land — ruled that voting computers used in the 2005 Bundestag election were unconstitutional. Not because anyone proved fraud. Because nobody could prove the absence of fraud. The judges said something that should be printed on the wall of every election commission office in the world: elections have to be checkable by an ordinary citizen, with no technical training, at every essential step. A machine that only a specialist can audit fails that test by definition, whether or not it's ever actually been tampered with. Germany went back to paper. It never went back to machines.

The Dutch didn't need a courtroom. They needed a television camera. In 2006, a group of hackers calling themselves "We do not trust voting computers" got hold of the same Nedap machines used across the country, reprogrammed one live on air, and — just to make the point unmistakable — taught it to play chess. If a voting machine can be turned into a chess computer by outsiders, what exactly do you think it can be turned into by insiders? The Dutch government's own commission confirmed the findings. Every Nedap machine in the country was decertified within a year.

Ireland is the one that should really sting, because it's just about money and stubbornness. Dublin spent roughly €51 million on electronic voting machines starting in 2002. An independent commission looked at the system and said, flatly, that it was "effectively self-auditing" — meaning nobody outside the machine could actually check its own math. Security experts hired by that same commission found it "very easy" to break in and take total control of the vote count. The government kept the machines in storage for years anyway, too embarrassed to admit the mistake, racking up millions more in warehouse fees. In 2009 they finally gave up. In 2012 they sold the entire fleet for scrap — about €70,000 for machines that cost €51 million. Not one of those machines ever counted a single real vote.

France put a moratorium on expanding voting machines in 2008 and has renewed it every few years since, most recently in 2022. Norway ran internet voting trials, had independent computer scientists review the code, got back a report describing "significant problems with coding style, security and correctness," and shut the whole programme down in 2014. Finland's supreme court actually annulled an election result in 2009 after an e-voting pilot lost votes and ordered a re-run on paper.

Notice the pattern? None of these are poor countries. None of these are technologically backward countries. These are places that could have thrown money at the problem until it went away. They didn't, because the problem isn't money. The problem is that a voting machine is, structurally, a black box — and you cannot build public trust on top of a black box, no matter how well-engineered it is.

The engineer who proved it — and got arrested for his trouble

In 2010, an Indian security researcher named Hari Prasad, working with University of Michigan computer scientist J. Alex Halderman and Dutch hacker Rop Gonggrijp — the same Gonggrijp who'd already helped take down the Dutch machines — got their hands on an actual Election Commission EVM. Not a demo unit. A real one, in use in Indian elections.

They didn't have the machine's source code — it's fused permanently into the chip, so they could only study it from the outside. Even working within that limitation, they demonstrated two separate ways to compromise it. First, they built a look-alike replacement for the display component, wired with hidden circuitry that could be preset to reallocate votes to a chosen candidate regardless of which button a voter actually pressed. Second, they built a small clip-on device that could read — and alter — the vote tallies stored in the machine's memory chip after polling had closed. Two different attacks, both requiring physical hands-on access to the hardware, both published, both demonstrated on video, both peer-reviewed.

And how did the Election Commission respond? Not with a technical rebuttal. With an arrest. Days after the Chief Election Commissioner told reporters the machines were "practically totally tamper proof," police showed up at Hari Prasad's house before dawn and drove him fourteen hours to Mumbai over a "theft" complaint about the machine he'd studied — a machine that had, by his account, simply been returned to where it came from. He spent eight days in custody. The court that granted him bail found the theft case didn't hold up. No charge-sheet was ever filed. That same year, the Electronic Frontier Foundation gave him an award for the work. The Indian state's answer to a documented security finding was a police van.

That should tell you everything about how seriously "tamper-proof" is meant to be taken as a claim versus how seriously it's meant to be believed.

"Nobody could hack it at our challenge" — well, of course not

You've probably heard the counter-argument: the Election Commission has repeatedly invited political parties to try and hack the EVMs, and nobody managed it. Case closed, right?

Look at what the "challenge" actually allows before you accept that. Participants get a fixed, short window — hours, not the weeks or months real researchers like Prasad's team took. They don't get the source code. They're not allowed to open the machine down to the chip level or attempt the kind of physical hardware substitution that was the whole point of the 2010 demonstration. The rules are written by the very body whose machines are being tested. When the 2017 challenge finally happened, only two parties out of dozens invited even showed up to try, and both of them spent their time looking at circuit boards rather than attempting anything resembling a real attack.

That's not a security audit. That's a guided tour with the word "challenge" stapled on. A genuine adversarial test gives independent researchers unrestricted access to hardware and code, on their own timeline, with no institutional referee standing over their shoulder deciding what counts as fair play. India has never offered that. Not once, in twenty years of this argument.

VVPAT was supposed to fix this. Did it?

The Voter Verifiable Paper Audit Trail was the Election Commission's answer to all of the above — a paper slip, generated alongside your electronic vote, that you can supposedly check for yourself. On paper — pun very much intended — this should have closed the trust gap for good.

But look at how it's actually implemented. You get seven seconds to look at a slip through a glass window before it drops into a box you'll never see opened unless your polling station happens to be one of a tiny, randomly selected handful that get manually counted. Since 2017 that viewing glass has been a dark, tinted one rather than a clear one — a change activists including the Right to Recall Party have pointed to as evidence that even the paper trail wasn't designed to make casual, ongoing observation easy. I want to be straight with you here: that specific group's claim that the dark glass is being used to actively swap or reuse slips undetected is their own theory, self-published, and it hasn't been independently verified by security researchers the way the 2010 hardware findings were — so take it as an allegation, not an established fact. But you don't need an unverified conspiracy theory to see the real problem sitting right there in plain sight: a "verifiable" paper trail that almost nobody's vote is ever actually verified against isn't doing the verifying. It's doing the reassuring. Those are not the same thing.

The movement India keeps trying to ignore

This isn't a fringe complaint from one bitter engineer. Over the years, the Socialist Party of India has argued, correctly, that paper fraud — the kind you'd get from old-style ballot stuffing — can at least be caught on camera, while electronic manipulation, if it happens, leaves nothing for anyone to see. The Right to Recall Party has spent years running public demonstrations trying to get ordinary voters to look harder at what they're being asked to trust. Groups have organised under banners like "EVM Hatao" — get rid of the EVM — collecting signatures and running street campaigns constituency by constituency. And across the political spectrum, at one point or another, leaders from nearly every major party — BJP included, back when they were in opposition after 2009 — have stood up and demanded the same thing: show us the paper.

That's not a coincidence. That's what happens when a system's opacity outlasts any single election cycle or any single party's convenience.

Give me a ballot paper I can watch with my own eyes

Nobody's pretending paper ballots are magic. Booth capturing happened. Ballots got stuffed. Counting took forever. Those were real problems, and EVMs did genuinely fix some of them.

But here's the difference that actually matters: when something goes wrong with a ballot paper, it's a visible crime. A stuffed box can be spotted, filmed, challenged, recounted by hand in front of party agents from every side of the political fight, right there in the room. When something goes wrong — or could go wrong — with a chip inside a sealed plastic case, none of that is available to you. Not to you, not to the candidate's polling agent, not to the journalist standing outside, not to the judge who might later be asked to rule on it. Booth capturing is a crime you can catch. A compromised control unit is a crime you might never even know to look for.

India already trusts paper for postal ballots. Servicemen, election officials on duty, people voting from outside their constituency — their votes go on paper, and nobody's demanding they switch to a machine for "efficiency." If paper is trustworthy enough for them, it's trustworthy enough for the other billion of us.

Germany didn't go back to paper because it's cheap or fast. It went back because paper is the one thing every voter, rich or poor, urban or rural, engineer or farmer, can watch with their own two eyes and understand without anybody's help. That's not a downgrade. That's the entire point of a democratic vote — that the person casting it doesn't have to trust anyone else's word for what happened to it.

Seven seconds behind a dark glass window isn't verification. It's a magic trick. And it's long past time we stopped applauding.

Watch these video to learn more about EVM manipulation :

1) https://youtu.be/wGmCTki8wqE?si=xF6yB91UsRsauOy4 ( in hindi ) 

2) https://youtu.be/ZlCOj1dElDY?si=KvwymRhr-288P6Qj ( in english ) 

3) https://youtu.be/VIBxCkxnOfo?feature=shared

4) https://youtu.be/s-vLbdVFefc?feature=shared

5)  https://youtu.be/AzGzxgn6_rA?si=I-z8wlOvRyCf656O