Saturday, June 6, 2009

Some technological issues with elections.

This post was long due given that it stems from an experience about 2 months ago. Whenever I need to understand something I go to a person who seems knowledgeable in that subject and talk to him/her. In this regard I have had a chat about various technological issues asssociated to voting, with a Theoretical Computer Science prof at TIFR (my current instutute) and in general my experience has been that people in the Theoretical Computer Science do have an amazing level of clarity with thinking and precision about everything they do.
I have a great admiration for people in this field.

So this prof is very young. May be just 5 years older to me. He recently got his PhD in the subject from UIUC (USA).He is a Keralite and did his B.Tech from IITKgp.

So here are the few insights that we generated during the discussion.

It isn't very clear that the use of EVMs has really made the voting process very significantly fair. Now it is just a matter of what kind of forces the political parties need to harness to rig the elections. Earlier it used to be local goons and thugs and now it probably has to be some electronics engineers and techie people. We have often joked about the possibility of the political parties going for "campussing" in the IITs and other such engineering institutes hiring tech people.

Earlier we have heard of such things happen in West Bengal: that therewould be a hole in on the top of the election room and a person would be sitting on the roof watching through the hole as to whom the person inside voted! This guy would be employed by the party X and if the person inside the booth didn't vote for part X then he would give a signal to the thugs outside from the roof-top , so that as soon as the person comes out he/she gets bashed up the goons.

This is where EVMs have definitely made a difference. "Secret Ballot" has become a little more secret than usual. But it hasn't probably made mcuh of a difference to the rigging situation. Thre have anyway been n number of claims of the EVM circuit being tweaked by some party the night before in the godown where the EVMs are kept.

Now the issue is what about the problem of people not being able to vote because of not being present in their constituency during their voting. This rule definitely needs to be changed given that such a large section of the youth voters live in the hostels of the various colleges far away from their home-towns. It would be a huge transportation chaos if all of the students have to be come home during the election date. They can have exams and what not. So it seems that there is good enough reason to start electronic voting and sms-voting.

The obvious issue with electronic voting is that of online identity resolution. (and there are slightly more non-obvious problems which I shall come on later) But then it isn't very clear why simple solutions like a password protected site with Captcha will not solve the problem. Only glitch being that since we have anyway experience of the "great" internet abilities of the Indian government it might be a good idea to lease out the running of these web-pages and voting sites to professional web-development companies.

And with sms voting the identity resolution seems to be slightly at stake unless we can develop some very sophisticated phones. In the common cell-phoes it is easily possible for a same person to cast multiple sms votes and I am not sure how password protected voting will work on cell-phones. Sms voting reminds me of the consistent fraud which hapenned with these reality-music shows which used sms-voting. Where the parents of the contestants distributed money in their localities to motivate people to send multiple smses in favour of their kids.

Now if we can develop very sophisticated phones which have say an iris pattern recognition system built on them which will get linked to the voting website as soon as the person logs into it. So the website will allow only 1 vote per iris pattern. But phones with such technologies to get into the market will probably take 50 more years.

Now consider the slightly non-obvious problem with internet voting. Imagine a person with a laptop going into a locality with a gun in his hand and asking all the people to vote for party X on the website by logging in to the web-site infront of him. Done! End of all election systems. With internet voting the "secretness" of the ballot will probably be totally gone. But then the issue of bribing the electorate for getting votes becomes a little more lucrative for the parties since now one can "see" whom the guy voted for. Now the parties can start "paying" for the votes only when the vote has been cast *infront of them*! The guy can come with a suitcase of cash and a laptop and a gun and sit with it and ask the people to queue infront of him. And then ask them to come up one by one and cast their vote in the net infront of him so that he is sure that they have voted for his party and then he will hand over the 1000Rs to the person.

So with internet voting there is greater "incentive" for the corrupt parties to bribe the electorate while the guarantee of return on the money used in bribing was less with EVM or the paper ballot. In EVMs and paper ballot the party has to pay the voter *before* the casting of the vote and there is no way for the party to check whether the bribed guy double crossed him! Hence with internet voting things can literally be "cash for vote" (With verification!). Hence a simple internet voting jeopardizes the sitaution further by giving greater incentive to corruption.


But a slight change in the mechanism of the voting web-site one can probably gain greater fairness with all the advantages of internet voting kept intact: By using something which we called the "last vote registering". Here the voter is allowed to vote multiple number of times and only his/her last vote will be counted for. So what happens is that even if the voter is forced to vote for a certain party infront of the man with the guna and cash he/she can still go back home or later sneak into a cyber-cafe somehwre and cast his vote again. Here the corrupt party loses all incentive for bribing since what this bribed guy votes infront of him can be completely reversed later!

But there will still be a way out for the really desperate and corrupt party. The corrupt party can ensure that except for the places where it has his men with laptop guns and cash at all other places the network is snapped of. So that all homes and cyber-cafes in the area don't have internet connection on that day. But even such a desperate attempt will not be able to deter the dedicated voter since she/he can stil vote *again* by going to her/his office which is likely to be far away from her/his locality and have an interbnet connection.

To end it my primary concern being not the fairness of the election process but the lives of the election officers lost in the election process in India. I consider them to be "martyrs" at the same level of respect as soldiers who die in the battlefield defending the nation. They fought for democracy and lost their lives.

It is a matter of great shame that even in this recent election in India so many election officers lost their lives.
My interests in looking at newer technologies of voting is to eliminate this concept of "election officers".