...and it doesn't address one of the key issues with votes being cast in places other than a booth in a polling station where you are alone: Coercion.

What's to stop a family member (OK, not much of an issue for overseas military personnel, one would think) or another person from hovering just outside the camera's capture area with a blunt object, ready to -ahem- influence your vote?

Granted, this would only affect individual votes; there are bigger potential issues with electronically tallying votes, but I'll leave those be...

> What's to stop a family member...[from] influenc[ing] your vote?

This is also a problem with mail-in ballots. (It’s why I believe early votes should be allowed to be overwritten by in-person votes.)

The issue with smartphone voting with selfie verification is almost everyone’s phones contain selfies. (Americans overseas are also likely to have a photo ID on their camera roll.) So controlling the phone means controlling a vote. That’s a scalable risk.

Moreover, if the app is anonymising after verifying, it means the app developers have been deputised by the Board of Elections to verify voters and count their votes. (Whether you voted is public, in America. For whom you voted is not.) That’s a big deal!

The blockchain add-on is worthless if done right, dangerous or illegal if done wrong, and increases the cross-section across which the above issues get amplified.

>This is also a problem with mail-in ballots.

Just as a datapoint this doesn't exist everywhere. In France for instance it's not possible to vote without physically going to a polling place unless you're living in a foreign country and only for certain elections (not possible for the presidential elections for instance). It is however possible to give a "procuration" for somebody else to vote for you but you still need to go to the police before the vote to make the procuration so it wouldn't be a very convenient way to fraud since you'd have to coerce the person into giving you their vote in front of a police officer.

I agree that if your electoral system accepts mail-in ballots then other electronic ways to vote remotely don't seem that much worse overall since they would share pretty much the same potential for abuse. I guess paper ballots would make it a bit harder to automate fraud at least.

Just to point out that "early voting" can take several forms, since I believe in the US many states combine (erroneously) mail-in voting with early voting.

In Australia, early voting can be done at polling places for several weeks before the election (though there's less polling places open obviously). Mail voting is also provided, but I agree with you that it really is something that has questionable merit -- when you're overseas you can vote at the Australian embassy. The only real use-case it serves is people who live in the outback or are worried about personal safety (who are "silent electors" -- where your address is not posted on the public voter roll), and I think that mail voting should only be provided to those people.

> It’s why I believe early votes should be allowed to be overwritten by in-person votes.

It's an interesting idea, though how would you maintain the secret ballot? If you ask them to make a "negative vote" they can suddenly vote three times without any chance of being caught.

You could keep the ballot sealed but tied to the voter's ID. Once the polling center closes for every mailed-in ballot they check if they have an in-person vote. If so, they discard it, otherwise they add it in the urn before it's opened for counting.

This is actually how it works in Sweden. They check the voter list whether the person already voted during the day and discards the mail-in if so.

> Once the polling center closes for every mailed-in ballot they check if they have an in-person vote

Agree on keeping votes sealed until live polling closes. But there is no need to check every mailed-in ballot against in-person votes.

Instead, when a mailed-in ballot arrives one marks the voter rolls. When a voter checks into a polling place, the rolls are anyway consulted. At that time, if a live voter shows as having already voted: that ballot is found and destroyed, and the voter given a fresh ballot to cast.

If you could re-vote as many times as you’d like, then this would largely destroy coercion since the coercer would have to control you or your device during the rest of the voting period. That alone would significantly reduce this risk. For more complex scenarios, you could allow the user to enter his/her PIN backwards which would mean “silently ignore this vote.”

>and “blockchain,” which it claims allows it to secure the voting process...

(I'm rubbing my temples right now)

I don't even have the words for how bad this would be for a democracy or republic.

I will, however, make a curt but general comment on blockchain:

Anytime you see that buzzword, be on your guard.

Question: Are there (m)any people other than those involved in the manufacture and sales of voting machines, or their lobbyists, that support increased automation in the voting process?

I'm absolutely, 100% happy with putting a pencil mark onto a paper ballot, which I fold and place into a box (with people watching), which is later taken (with people watching) to a counting place, emptied out on a table (with people watching), and counted by hand by volunteers (with yet more people watching).

Pen and paper is the objectively superior technology for the job; it's the Minimum Viable Product that is auditable by the most number of people. You shouldn't need to be a crypto expert to be able to confirm that your vote was counted. Public trust should be an explicit design goal, and arguably the most important one.

I'm with you on that but I was surprised to find out during the last Brazilian election that they seem to do most of their voting on electronic machines and it seems well accepted over there. At least when I mentioned that to Brazilian colleagues they seemed fine with it.

This video (in Portuguese) shows a "tear down" endorsed by the manufacturer: https://www.youtube.com/watch?v=4wrMLzqgKEI

Both the tone of the video (unsurprisingly, given its "officially endorsed" nature) but also the comments seem overall pretty positive. Even though they mention that the software is closed source and you basically have to trust the manufacturer (and any third parties making hardware and software for the device along the way).

Politicians, political parties, and governments (from local to national).

For some reason the lure of instant results is so great that every year there is some place in the world where electronic voting is once again strongly pushed for, even though none of the arguments against it are suitably addressed.

Is it a mindset thing? A disdain for an analogue process that they deem outmoded? A lack of understanding of the inherent problems with automatization for this particular task? All I know is that it takes a lot of effort to keep elections free from voting machines.

This is a global phenomenon by the way.

For the average person, it probably seems weird that you can buy groceries online but you have to go to a place with a special machine to vote. Online voting seems like an obvious win, if you don't realize how fragile online security is.

Finally.

REASON!

Now this guy has the right idea.

Why are people trying to fix something that is not broken?

A good question which could also be asked of 30% of HN articles

Don't worry, I'm gonna solve the trust problem in blockchain technology by creating a blockchain that lets you know which of the blockchains are trustworthy. It will be fully secured by blockchain technology, so you know you can trust it.

Honestly, I don’t think voting by app is inherently a “horrifically bad idea” – just that Voatz’ implementation is horrifically bad.

Biometrics are going to be almost impossible to do in a non-invasive and secure way for now, so cut that part of the pitch. Blockchain may not be an entirely round peg in this square hole, but there are other cryptographic architectures that probably make more sense (I am not a cryptographer, but maybe go talk to a competent one before just deciding on blockchain).

Obviously, without the two big buzzwords of blockchain and biometrics, you’re suddenly at a big disadvantage for VC money, but if mobile voting is ever going to be feasible, it’s going to need a considered, end-to-end approach that isn’t some MVP built from off-the-shelf components and off-the-shelf ideas.

> I don’t think voting by app is inherently a “horrifically bad idea”

If a voter can cast their vote from their phone, anyone else can.

The strength of paper-based voting is in its cost and complexity: the number of confederates required to tamper with an election is prohibitively large. With app-based voting comes the risk of a large number of compromised phones tampering with elections.

The security downsides simply don't make sense next to mailed-in ballots, which capture the security (and accessibility) of paper ballots with the convenience of voting from home. The only upside of app-based voting is cost, which is a ridiculous factor to strongly optimize for when it comes to something as important as elections.

[1] https://en.wikipedia.org/wiki/United_States_presidential_ele...

> Honestly, I don’t think voting by app is inherently a “horrifically bad idea”

Yes, it is. There are many ways such a system can go wrong. If its not this defective implementation, it'll be some other company's defective implementation. If not that, it'll be a targeted attack against the underlying platform.

Another big thing you have to think about is... Who will be choosing the system? Its going to be some bureaucrats with a vested interest in supporting one particular vendor, who use their power to silence any critics of said vendor's solution. Whenever electronic voting comes up, we always end up with this situation.

What's better about traditional methods? Numbers are always off, recounts always get different results. It still seems like a black box.

So who's betting both the "anonymized" votes and the biometric software are both going to end up having time-stamped logs, which we'll only discover after a major election takes place?

I wonder how many pairs of ID pic and selfie video can directly be grabbed off Instagram.

Apparently there is an eerily relevant XKCD: https://xkcd.com/2030/